Model save
Browse files- README.md +61 -0
- adapter_model.safetensors +1 -1
- all_results.json +8 -0
- runs/May16_02-42-49_gpu4-119-5/events.out.tfevents.1715791725.gpu4-119-5.3642792.0 +2 -2
- train_results.json +8 -0
- trainer_state.json +2256 -0
README.md
ADDED
|
@@ -0,0 +1,61 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
library_name: peft
|
| 3 |
+
tags:
|
| 4 |
+
- trl
|
| 5 |
+
- dpo
|
| 6 |
+
- generated_from_trainer
|
| 7 |
+
base_model: DUAL-GPO/phi-2-sft-lora-ultrachat-merged
|
| 8 |
+
model-index:
|
| 9 |
+
- name: phi-2-gpo-v30-i1
|
| 10 |
+
results: []
|
| 11 |
+
---
|
| 12 |
+
|
| 13 |
+
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
| 14 |
+
should probably proofread and complete it, then remove this comment. -->
|
| 15 |
+
|
| 16 |
+
# phi-2-gpo-v30-i1
|
| 17 |
+
|
| 18 |
+
This model is a fine-tuned version of [DUAL-GPO/phi-2-sft-lora-ultrachat-merged](https://huggingface.co/DUAL-GPO/phi-2-sft-lora-ultrachat-merged) on the None dataset.
|
| 19 |
+
|
| 20 |
+
## Model description
|
| 21 |
+
|
| 22 |
+
More information needed
|
| 23 |
+
|
| 24 |
+
## Intended uses & limitations
|
| 25 |
+
|
| 26 |
+
More information needed
|
| 27 |
+
|
| 28 |
+
## Training and evaluation data
|
| 29 |
+
|
| 30 |
+
More information needed
|
| 31 |
+
|
| 32 |
+
## Training procedure
|
| 33 |
+
|
| 34 |
+
### Training hyperparameters
|
| 35 |
+
|
| 36 |
+
The following hyperparameters were used during training:
|
| 37 |
+
- learning_rate: 5e-06
|
| 38 |
+
- train_batch_size: 4
|
| 39 |
+
- eval_batch_size: 4
|
| 40 |
+
- seed: 42
|
| 41 |
+
- distributed_type: multi-GPU
|
| 42 |
+
- num_devices: 3
|
| 43 |
+
- gradient_accumulation_steps: 4
|
| 44 |
+
- total_train_batch_size: 48
|
| 45 |
+
- total_eval_batch_size: 12
|
| 46 |
+
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 47 |
+
- lr_scheduler_type: cosine
|
| 48 |
+
- lr_scheduler_warmup_ratio: 0.1
|
| 49 |
+
- num_epochs: 1
|
| 50 |
+
|
| 51 |
+
### Training results
|
| 52 |
+
|
| 53 |
+
|
| 54 |
+
|
| 55 |
+
### Framework versions
|
| 56 |
+
|
| 57 |
+
- PEFT 0.7.1
|
| 58 |
+
- Transformers 4.36.2
|
| 59 |
+
- Pytorch 2.1.2+cu121
|
| 60 |
+
- Datasets 2.14.6
|
| 61 |
+
- Tokenizers 0.15.2
|
adapter_model.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 167807296
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:510232fe7039f3cb44011d566a02229117759aabefd9a6c4bdd80afb41238df0
|
| 3 |
size 167807296
|
all_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 1.0,
|
| 3 |
+
"train_loss": 0.09621614837856943,
|
| 4 |
+
"train_runtime": 21999.9134,
|
| 5 |
+
"train_samples": 76135,
|
| 6 |
+
"train_samples_per_second": 3.461,
|
| 7 |
+
"train_steps_per_second": 0.072
|
| 8 |
+
}
|
runs/May16_02-42-49_gpu4-119-5/events.out.tfevents.1715791725.gpu4-119-5.3642792.0
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:88e65308ea443b9380226f387d602292f015163f93b51b3a15791dcb319a7e39
|
| 3 |
+
size 105897
|
train_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 1.0,
|
| 3 |
+
"train_loss": 0.09621614837856943,
|
| 4 |
+
"train_runtime": 21999.9134,
|
| 5 |
+
"train_samples": 76135,
|
| 6 |
+
"train_samples_per_second": 3.461,
|
| 7 |
+
"train_steps_per_second": 0.072
|
| 8 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,2256 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 0.9998423955870764,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 1586,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.0,
|
| 13 |
+
"learning_rate": 3.1446540880503146e-08,
|
| 14 |
+
"logits/chosen": 0.20935016870498657,
|
| 15 |
+
"logits/rejected": 0.28666022419929504,
|
| 16 |
+
"logps/chosen": -338.6566467285156,
|
| 17 |
+
"logps/rejected": -308.9756164550781,
|
| 18 |
+
"loss": 0.1265,
|
| 19 |
+
"rewards/accuracies": 0.0,
|
| 20 |
+
"rewards/chosen": 0.0,
|
| 21 |
+
"rewards/margins": 0.0,
|
| 22 |
+
"rewards/rejected": 0.0,
|
| 23 |
+
"step": 1
|
| 24 |
+
},
|
| 25 |
+
{
|
| 26 |
+
"epoch": 0.01,
|
| 27 |
+
"learning_rate": 3.144654088050315e-07,
|
| 28 |
+
"logits/chosen": 0.21327850222587585,
|
| 29 |
+
"logits/rejected": 0.2737487554550171,
|
| 30 |
+
"logps/chosen": -325.64306640625,
|
| 31 |
+
"logps/rejected": -319.2493896484375,
|
| 32 |
+
"loss": 0.1202,
|
| 33 |
+
"rewards/accuracies": 0.4027777910232544,
|
| 34 |
+
"rewards/chosen": 4.140691817156039e-05,
|
| 35 |
+
"rewards/margins": -2.3560824047308415e-05,
|
| 36 |
+
"rewards/rejected": 6.496775313280523e-05,
|
| 37 |
+
"step": 10
|
| 38 |
+
},
|
| 39 |
+
{
|
| 40 |
+
"epoch": 0.01,
|
| 41 |
+
"learning_rate": 6.28930817610063e-07,
|
| 42 |
+
"logits/chosen": 0.260469526052475,
|
| 43 |
+
"logits/rejected": 0.3009797930717468,
|
| 44 |
+
"logps/chosen": -294.43804931640625,
|
| 45 |
+
"logps/rejected": -289.858154296875,
|
| 46 |
+
"loss": 0.1078,
|
| 47 |
+
"rewards/accuracies": 0.44999998807907104,
|
| 48 |
+
"rewards/chosen": 3.346197263454087e-05,
|
| 49 |
+
"rewards/margins": 5.3948046115692705e-05,
|
| 50 |
+
"rewards/rejected": -2.0486073481151834e-05,
|
| 51 |
+
"step": 20
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.02,
|
| 55 |
+
"learning_rate": 9.433962264150944e-07,
|
| 56 |
+
"logits/chosen": 0.19516317546367645,
|
| 57 |
+
"logits/rejected": 0.22890889644622803,
|
| 58 |
+
"logps/chosen": -330.41143798828125,
|
| 59 |
+
"logps/rejected": -330.6087341308594,
|
| 60 |
+
"loss": 0.1203,
|
| 61 |
+
"rewards/accuracies": 0.4937500059604645,
|
| 62 |
+
"rewards/chosen": -5.2339473768370226e-05,
|
| 63 |
+
"rewards/margins": 0.000176291519892402,
|
| 64 |
+
"rewards/rejected": -0.0002286309900227934,
|
| 65 |
+
"step": 30
|
| 66 |
+
},
|
| 67 |
+
{
|
| 68 |
+
"epoch": 0.03,
|
| 69 |
+
"learning_rate": 1.257861635220126e-06,
|
| 70 |
+
"logits/chosen": 0.22363004088401794,
|
| 71 |
+
"logits/rejected": 0.2845616936683655,
|
| 72 |
+
"logps/chosen": -351.8126525878906,
|
| 73 |
+
"logps/rejected": -298.5321044921875,
|
| 74 |
+
"loss": 0.1223,
|
| 75 |
+
"rewards/accuracies": 0.5687500238418579,
|
| 76 |
+
"rewards/chosen": -0.0003051063104066998,
|
| 77 |
+
"rewards/margins": 0.0003100560570601374,
|
| 78 |
+
"rewards/rejected": -0.0006151623092591763,
|
| 79 |
+
"step": 40
|
| 80 |
+
},
|
| 81 |
+
{
|
| 82 |
+
"epoch": 0.03,
|
| 83 |
+
"learning_rate": 1.5723270440251573e-06,
|
| 84 |
+
"logits/chosen": 0.2019948959350586,
|
| 85 |
+
"logits/rejected": 0.25790125131607056,
|
| 86 |
+
"logps/chosen": -321.6785888671875,
|
| 87 |
+
"logps/rejected": -317.89813232421875,
|
| 88 |
+
"loss": 0.1163,
|
| 89 |
+
"rewards/accuracies": 0.5,
|
| 90 |
+
"rewards/chosen": -0.0004650927730835974,
|
| 91 |
+
"rewards/margins": 0.0002983523881994188,
|
| 92 |
+
"rewards/rejected": -0.0007634450448676944,
|
| 93 |
+
"step": 50
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 0.04,
|
| 97 |
+
"learning_rate": 1.8867924528301889e-06,
|
| 98 |
+
"logits/chosen": 0.23230671882629395,
|
| 99 |
+
"logits/rejected": 0.24606795608997345,
|
| 100 |
+
"logps/chosen": -320.841552734375,
|
| 101 |
+
"logps/rejected": -297.08306884765625,
|
| 102 |
+
"loss": 0.1075,
|
| 103 |
+
"rewards/accuracies": 0.512499988079071,
|
| 104 |
+
"rewards/chosen": -0.0008995781536214054,
|
| 105 |
+
"rewards/margins": 0.0004379089514259249,
|
| 106 |
+
"rewards/rejected": -0.0013374872505664825,
|
| 107 |
+
"step": 60
|
| 108 |
+
},
|
| 109 |
+
{
|
| 110 |
+
"epoch": 0.04,
|
| 111 |
+
"learning_rate": 2.2012578616352204e-06,
|
| 112 |
+
"logits/chosen": 0.18308493494987488,
|
| 113 |
+
"logits/rejected": 0.2341008186340332,
|
| 114 |
+
"logps/chosen": -288.20172119140625,
|
| 115 |
+
"logps/rejected": -298.20050048828125,
|
| 116 |
+
"loss": 0.1169,
|
| 117 |
+
"rewards/accuracies": 0.512499988079071,
|
| 118 |
+
"rewards/chosen": -0.0013604220002889633,
|
| 119 |
+
"rewards/margins": 0.0006924796034581959,
|
| 120 |
+
"rewards/rejected": -0.0020529015455394983,
|
| 121 |
+
"step": 70
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"epoch": 0.05,
|
| 125 |
+
"learning_rate": 2.515723270440252e-06,
|
| 126 |
+
"logits/chosen": 0.24107146263122559,
|
| 127 |
+
"logits/rejected": 0.2765248417854309,
|
| 128 |
+
"logps/chosen": -327.1878356933594,
|
| 129 |
+
"logps/rejected": -324.80706787109375,
|
| 130 |
+
"loss": 0.1138,
|
| 131 |
+
"rewards/accuracies": 0.48124998807907104,
|
| 132 |
+
"rewards/chosen": -0.0024016976822167635,
|
| 133 |
+
"rewards/margins": 0.0004348896036390215,
|
| 134 |
+
"rewards/rejected": -0.0028365873731672764,
|
| 135 |
+
"step": 80
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 0.06,
|
| 139 |
+
"learning_rate": 2.830188679245283e-06,
|
| 140 |
+
"logits/chosen": 0.23600873351097107,
|
| 141 |
+
"logits/rejected": 0.2604605257511139,
|
| 142 |
+
"logps/chosen": -322.81610107421875,
|
| 143 |
+
"logps/rejected": -312.8374328613281,
|
| 144 |
+
"loss": 0.1158,
|
| 145 |
+
"rewards/accuracies": 0.5249999761581421,
|
| 146 |
+
"rewards/chosen": -0.0036561130546033382,
|
| 147 |
+
"rewards/margins": 0.0011059448588639498,
|
| 148 |
+
"rewards/rejected": -0.004762057680636644,
|
| 149 |
+
"step": 90
|
| 150 |
+
},
|
| 151 |
+
{
|
| 152 |
+
"epoch": 0.06,
|
| 153 |
+
"learning_rate": 3.1446540880503146e-06,
|
| 154 |
+
"logits/chosen": 0.17113199830055237,
|
| 155 |
+
"logits/rejected": 0.19848307967185974,
|
| 156 |
+
"logps/chosen": -339.1964416503906,
|
| 157 |
+
"logps/rejected": -318.428955078125,
|
| 158 |
+
"loss": 0.1188,
|
| 159 |
+
"rewards/accuracies": 0.512499988079071,
|
| 160 |
+
"rewards/chosen": -0.0050820643082261086,
|
| 161 |
+
"rewards/margins": 0.002656476106494665,
|
| 162 |
+
"rewards/rejected": -0.007738540414720774,
|
| 163 |
+
"step": 100
|
| 164 |
+
},
|
| 165 |
+
{
|
| 166 |
+
"epoch": 0.07,
|
| 167 |
+
"learning_rate": 3.4591194968553458e-06,
|
| 168 |
+
"logits/chosen": 0.16253027319908142,
|
| 169 |
+
"logits/rejected": 0.20139364898204803,
|
| 170 |
+
"logps/chosen": -339.23040771484375,
|
| 171 |
+
"logps/rejected": -336.61444091796875,
|
| 172 |
+
"loss": 0.1205,
|
| 173 |
+
"rewards/accuracies": 0.5375000238418579,
|
| 174 |
+
"rewards/chosen": -0.008081575855612755,
|
| 175 |
+
"rewards/margins": 0.0035987249575555325,
|
| 176 |
+
"rewards/rejected": -0.011680299416184425,
|
| 177 |
+
"step": 110
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 0.08,
|
| 181 |
+
"learning_rate": 3.7735849056603777e-06,
|
| 182 |
+
"logits/chosen": 0.17907488346099854,
|
| 183 |
+
"logits/rejected": 0.1907232403755188,
|
| 184 |
+
"logps/chosen": -347.43646240234375,
|
| 185 |
+
"logps/rejected": -350.59881591796875,
|
| 186 |
+
"loss": 0.1156,
|
| 187 |
+
"rewards/accuracies": 0.625,
|
| 188 |
+
"rewards/chosen": -0.012615310959517956,
|
| 189 |
+
"rewards/margins": 0.0076774852350354195,
|
| 190 |
+
"rewards/rejected": -0.020292794331908226,
|
| 191 |
+
"step": 120
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"epoch": 0.08,
|
| 195 |
+
"learning_rate": 4.08805031446541e-06,
|
| 196 |
+
"logits/chosen": 0.14514799416065216,
|
| 197 |
+
"logits/rejected": 0.1459697037935257,
|
| 198 |
+
"logps/chosen": -318.0201416015625,
|
| 199 |
+
"logps/rejected": -335.25506591796875,
|
| 200 |
+
"loss": 0.1111,
|
| 201 |
+
"rewards/accuracies": 0.59375,
|
| 202 |
+
"rewards/chosen": -0.01840413734316826,
|
| 203 |
+
"rewards/margins": 0.015484845265746117,
|
| 204 |
+
"rewards/rejected": -0.033888980746269226,
|
| 205 |
+
"step": 130
|
| 206 |
+
},
|
| 207 |
+
{
|
| 208 |
+
"epoch": 0.09,
|
| 209 |
+
"learning_rate": 4.402515723270441e-06,
|
| 210 |
+
"logits/chosen": 0.10525090992450714,
|
| 211 |
+
"logits/rejected": 0.12337298691272736,
|
| 212 |
+
"logps/chosen": -353.93353271484375,
|
| 213 |
+
"logps/rejected": -395.66522216796875,
|
| 214 |
+
"loss": 0.117,
|
| 215 |
+
"rewards/accuracies": 0.59375,
|
| 216 |
+
"rewards/chosen": -0.026070892810821533,
|
| 217 |
+
"rewards/margins": 0.020400088280439377,
|
| 218 |
+
"rewards/rejected": -0.04647097736597061,
|
| 219 |
+
"step": 140
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 0.09,
|
| 223 |
+
"learning_rate": 4.716981132075472e-06,
|
| 224 |
+
"logits/chosen": 0.104182168841362,
|
| 225 |
+
"logits/rejected": 0.09100546687841415,
|
| 226 |
+
"logps/chosen": -375.31732177734375,
|
| 227 |
+
"logps/rejected": -394.6401062011719,
|
| 228 |
+
"loss": 0.1104,
|
| 229 |
+
"rewards/accuracies": 0.612500011920929,
|
| 230 |
+
"rewards/chosen": -0.04359036684036255,
|
| 231 |
+
"rewards/margins": 0.019685441628098488,
|
| 232 |
+
"rewards/rejected": -0.06327580660581589,
|
| 233 |
+
"step": 150
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"epoch": 0.1,
|
| 237 |
+
"learning_rate": 4.999993941548484e-06,
|
| 238 |
+
"logits/chosen": 0.06796596199274063,
|
| 239 |
+
"logits/rejected": 0.10237328708171844,
|
| 240 |
+
"logps/chosen": -380.8462829589844,
|
| 241 |
+
"logps/rejected": -407.0494079589844,
|
| 242 |
+
"loss": 0.1092,
|
| 243 |
+
"rewards/accuracies": 0.5375000238418579,
|
| 244 |
+
"rewards/chosen": -0.05505872890353203,
|
| 245 |
+
"rewards/margins": 0.028139878064393997,
|
| 246 |
+
"rewards/rejected": -0.08319860696792603,
|
| 247 |
+
"step": 160
|
| 248 |
+
},
|
| 249 |
+
{
|
| 250 |
+
"epoch": 0.11,
|
| 251 |
+
"learning_rate": 4.999266962896092e-06,
|
| 252 |
+
"logits/chosen": 0.08151620626449585,
|
| 253 |
+
"logits/rejected": 0.07926828414201736,
|
| 254 |
+
"logps/chosen": -364.17974853515625,
|
| 255 |
+
"logps/rejected": -370.903564453125,
|
| 256 |
+
"loss": 0.1007,
|
| 257 |
+
"rewards/accuracies": 0.5625,
|
| 258 |
+
"rewards/chosen": -0.053988486528396606,
|
| 259 |
+
"rewards/margins": 0.021424416452646255,
|
| 260 |
+
"rewards/rejected": -0.07541291415691376,
|
| 261 |
+
"step": 170
|
| 262 |
+
},
|
| 263 |
+
{
|
| 264 |
+
"epoch": 0.11,
|
| 265 |
+
"learning_rate": 4.997328697661372e-06,
|
| 266 |
+
"logits/chosen": 0.0645218938589096,
|
| 267 |
+
"logits/rejected": 0.09341205656528473,
|
| 268 |
+
"logps/chosen": -343.18609619140625,
|
| 269 |
+
"logps/rejected": -353.36810302734375,
|
| 270 |
+
"loss": 0.1208,
|
| 271 |
+
"rewards/accuracies": 0.543749988079071,
|
| 272 |
+
"rewards/chosen": -0.049203019589185715,
|
| 273 |
+
"rewards/margins": 0.016046565026044846,
|
| 274 |
+
"rewards/rejected": -0.06524958461523056,
|
| 275 |
+
"step": 180
|
| 276 |
+
},
|
| 277 |
+
{
|
| 278 |
+
"epoch": 0.12,
|
| 279 |
+
"learning_rate": 4.994180085237635e-06,
|
| 280 |
+
"logits/chosen": 0.13387538492679596,
|
| 281 |
+
"logits/rejected": 0.13471734523773193,
|
| 282 |
+
"logps/chosen": -333.0649108886719,
|
| 283 |
+
"logps/rejected": -387.0137939453125,
|
| 284 |
+
"loss": 0.1179,
|
| 285 |
+
"rewards/accuracies": 0.5375000238418579,
|
| 286 |
+
"rewards/chosen": -0.05196870118379593,
|
| 287 |
+
"rewards/margins": 0.02883673831820488,
|
| 288 |
+
"rewards/rejected": -0.08080543577671051,
|
| 289 |
+
"step": 190
|
| 290 |
+
},
|
| 291 |
+
{
|
| 292 |
+
"epoch": 0.13,
|
| 293 |
+
"learning_rate": 4.98982265162112e-06,
|
| 294 |
+
"logits/chosen": 0.09480293095111847,
|
| 295 |
+
"logits/rejected": 0.08700738847255707,
|
| 296 |
+
"logps/chosen": -373.3707580566406,
|
| 297 |
+
"logps/rejected": -419.16131591796875,
|
| 298 |
+
"loss": 0.0986,
|
| 299 |
+
"rewards/accuracies": 0.5625,
|
| 300 |
+
"rewards/chosen": -0.058679305016994476,
|
| 301 |
+
"rewards/margins": 0.025583749637007713,
|
| 302 |
+
"rewards/rejected": -0.08426305651664734,
|
| 303 |
+
"step": 200
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"epoch": 0.13,
|
| 307 |
+
"learning_rate": 4.984258508671404e-06,
|
| 308 |
+
"logits/chosen": 0.12381438910961151,
|
| 309 |
+
"logits/rejected": 0.11012091487646103,
|
| 310 |
+
"logps/chosen": -403.3990478515625,
|
| 311 |
+
"logps/rejected": -435.538818359375,
|
| 312 |
+
"loss": 0.0994,
|
| 313 |
+
"rewards/accuracies": 0.625,
|
| 314 |
+
"rewards/chosen": -0.051610808819532394,
|
| 315 |
+
"rewards/margins": 0.041461896151304245,
|
| 316 |
+
"rewards/rejected": -0.09307270497083664,
|
| 317 |
+
"step": 210
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 0.14,
|
| 321 |
+
"learning_rate": 4.977490353087878e-06,
|
| 322 |
+
"logits/chosen": 0.13996745645999908,
|
| 323 |
+
"logits/rejected": 0.15085197985172272,
|
| 324 |
+
"logps/chosen": -365.3772277832031,
|
| 325 |
+
"logps/rejected": -394.3701477050781,
|
| 326 |
+
"loss": 0.0994,
|
| 327 |
+
"rewards/accuracies": 0.574999988079071,
|
| 328 |
+
"rewards/chosen": -0.046448659151792526,
|
| 329 |
+
"rewards/margins": 0.0269100870937109,
|
| 330 |
+
"rewards/rejected": -0.07335875183343887,
|
| 331 |
+
"step": 220
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 0.14,
|
| 335 |
+
"learning_rate": 4.969521465102775e-06,
|
| 336 |
+
"logits/chosen": 0.10541512072086334,
|
| 337 |
+
"logits/rejected": 0.13170066475868225,
|
| 338 |
+
"logps/chosen": -372.6682434082031,
|
| 339 |
+
"logps/rejected": -432.30670166015625,
|
| 340 |
+
"loss": 0.1162,
|
| 341 |
+
"rewards/accuracies": 0.5874999761581421,
|
| 342 |
+
"rewards/chosen": -0.055505841970443726,
|
| 343 |
+
"rewards/margins": 0.03007945977151394,
|
| 344 |
+
"rewards/rejected": -0.08558530360460281,
|
| 345 |
+
"step": 230
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 0.15,
|
| 349 |
+
"learning_rate": 4.960355706891375e-06,
|
| 350 |
+
"logits/chosen": 0.08097681403160095,
|
| 351 |
+
"logits/rejected": 0.08701089024543762,
|
| 352 |
+
"logps/chosen": -377.55023193359375,
|
| 353 |
+
"logps/rejected": -453.13702392578125,
|
| 354 |
+
"loss": 0.1107,
|
| 355 |
+
"rewards/accuracies": 0.6499999761581421,
|
| 356 |
+
"rewards/chosen": -0.061971623450517654,
|
| 357 |
+
"rewards/margins": 0.04781196638941765,
|
| 358 |
+
"rewards/rejected": -0.1097835898399353,
|
| 359 |
+
"step": 240
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 0.16,
|
| 363 |
+
"learning_rate": 4.94999752070019e-06,
|
| 364 |
+
"logits/chosen": 0.10144323110580444,
|
| 365 |
+
"logits/rejected": 0.13508056104183197,
|
| 366 |
+
"logps/chosen": -362.32012939453125,
|
| 367 |
+
"logps/rejected": -365.26947021484375,
|
| 368 |
+
"loss": 0.1084,
|
| 369 |
+
"rewards/accuracies": 0.543749988079071,
|
| 370 |
+
"rewards/chosen": -0.06062735244631767,
|
| 371 |
+
"rewards/margins": 0.024900587275624275,
|
| 372 |
+
"rewards/rejected": -0.0855279415845871,
|
| 373 |
+
"step": 250
|
| 374 |
+
},
|
| 375 |
+
{
|
| 376 |
+
"epoch": 0.16,
|
| 377 |
+
"learning_rate": 4.938451926693985e-06,
|
| 378 |
+
"logits/chosen": 0.11075524240732193,
|
| 379 |
+
"logits/rejected": 0.1454174965620041,
|
| 380 |
+
"logps/chosen": -375.95538330078125,
|
| 381 |
+
"logps/rejected": -407.412109375,
|
| 382 |
+
"loss": 0.1061,
|
| 383 |
+
"rewards/accuracies": 0.574999988079071,
|
| 384 |
+
"rewards/chosen": -0.07085081189870834,
|
| 385 |
+
"rewards/margins": 0.03195417672395706,
|
| 386 |
+
"rewards/rejected": -0.10280498117208481,
|
| 387 |
+
"step": 260
|
| 388 |
+
},
|
| 389 |
+
{
|
| 390 |
+
"epoch": 0.17,
|
| 391 |
+
"learning_rate": 4.9257245205227296e-06,
|
| 392 |
+
"logits/chosen": 0.14314480125904083,
|
| 393 |
+
"logits/rejected": 0.122921422123909,
|
| 394 |
+
"logps/chosen": -381.03472900390625,
|
| 395 |
+
"logps/rejected": -460.407470703125,
|
| 396 |
+
"loss": 0.0954,
|
| 397 |
+
"rewards/accuracies": 0.65625,
|
| 398 |
+
"rewards/chosen": -0.06635353714227676,
|
| 399 |
+
"rewards/margins": 0.05833283066749573,
|
| 400 |
+
"rewards/rejected": -0.12468637526035309,
|
| 401 |
+
"step": 270
|
| 402 |
+
},
|
| 403 |
+
{
|
| 404 |
+
"epoch": 0.18,
|
| 405 |
+
"learning_rate": 4.911821470609627e-06,
|
| 406 |
+
"logits/chosen": 0.11854016780853271,
|
| 407 |
+
"logits/rejected": 0.0927438884973526,
|
| 408 |
+
"logps/chosen": -393.5327453613281,
|
| 409 |
+
"logps/rejected": -386.33795166015625,
|
| 410 |
+
"loss": 0.1045,
|
| 411 |
+
"rewards/accuracies": 0.5687500238418579,
|
| 412 |
+
"rewards/chosen": -0.08982465416193008,
|
| 413 |
+
"rewards/margins": 0.011397195979952812,
|
| 414 |
+
"rewards/rejected": -0.10122185945510864,
|
| 415 |
+
"step": 280
|
| 416 |
+
},
|
| 417 |
+
{
|
| 418 |
+
"epoch": 0.18,
|
| 419 |
+
"learning_rate": 4.896749515161544e-06,
|
| 420 |
+
"logits/chosen": 0.16842451691627502,
|
| 421 |
+
"logits/rejected": 0.13145343959331512,
|
| 422 |
+
"logps/chosen": -371.6466064453125,
|
| 423 |
+
"logps/rejected": -430.08587646484375,
|
| 424 |
+
"loss": 0.0924,
|
| 425 |
+
"rewards/accuracies": 0.6000000238418579,
|
| 426 |
+
"rewards/chosen": -0.06959854066371918,
|
| 427 |
+
"rewards/margins": 0.037486035376787186,
|
| 428 |
+
"rewards/rejected": -0.10708458721637726,
|
| 429 |
+
"step": 290
|
| 430 |
+
},
|
| 431 |
+
{
|
| 432 |
+
"epoch": 0.19,
|
| 433 |
+
"learning_rate": 4.880515958903296e-06,
|
| 434 |
+
"logits/chosen": 0.14044585824012756,
|
| 435 |
+
"logits/rejected": 0.14415714144706726,
|
| 436 |
+
"logps/chosen": -387.0291442871094,
|
| 437 |
+
"logps/rejected": -432.3477478027344,
|
| 438 |
+
"loss": 0.0995,
|
| 439 |
+
"rewards/accuracies": 0.606249988079071,
|
| 440 |
+
"rewards/chosen": -0.07424234598875046,
|
| 441 |
+
"rewards/margins": 0.029624024406075478,
|
| 442 |
+
"rewards/rejected": -0.10386637598276138,
|
| 443 |
+
"step": 300
|
| 444 |
+
},
|
| 445 |
+
{
|
| 446 |
+
"epoch": 0.2,
|
| 447 |
+
"learning_rate": 4.8631286695373626e-06,
|
| 448 |
+
"logits/chosen": 0.12493617832660675,
|
| 449 |
+
"logits/rejected": 0.16041526198387146,
|
| 450 |
+
"logps/chosen": -432.5501403808594,
|
| 451 |
+
"logps/rejected": -449.15167236328125,
|
| 452 |
+
"loss": 0.0916,
|
| 453 |
+
"rewards/accuracies": 0.5062500238418579,
|
| 454 |
+
"rewards/chosen": -0.08585993945598602,
|
| 455 |
+
"rewards/margins": 0.024735430255532265,
|
| 456 |
+
"rewards/rejected": -0.11059536784887314,
|
| 457 |
+
"step": 310
|
| 458 |
+
},
|
| 459 |
+
{
|
| 460 |
+
"epoch": 0.2,
|
| 461 |
+
"learning_rate": 4.844596073930749e-06,
|
| 462 |
+
"logits/chosen": 0.11006021499633789,
|
| 463 |
+
"logits/rejected": 0.15848229825496674,
|
| 464 |
+
"logps/chosen": -401.59051513671875,
|
| 465 |
+
"logps/rejected": -428.4137268066406,
|
| 466 |
+
"loss": 0.1124,
|
| 467 |
+
"rewards/accuracies": 0.574999988079071,
|
| 468 |
+
"rewards/chosen": -0.08783275634050369,
|
| 469 |
+
"rewards/margins": 0.04196416586637497,
|
| 470 |
+
"rewards/rejected": -0.12979692220687866,
|
| 471 |
+
"step": 320
|
| 472 |
+
},
|
| 473 |
+
{
|
| 474 |
+
"epoch": 0.21,
|
| 475 |
+
"learning_rate": 4.824927154030856e-06,
|
| 476 |
+
"logits/chosen": 0.14780564606189728,
|
| 477 |
+
"logits/rejected": 0.1997770518064499,
|
| 478 |
+
"logps/chosen": -412.12664794921875,
|
| 479 |
+
"logps/rejected": -425.85711669921875,
|
| 480 |
+
"loss": 0.1016,
|
| 481 |
+
"rewards/accuracies": 0.5375000238418579,
|
| 482 |
+
"rewards/chosen": -0.08760993927717209,
|
| 483 |
+
"rewards/margins": 0.03202151134610176,
|
| 484 |
+
"rewards/rejected": -0.11963143199682236,
|
| 485 |
+
"step": 330
|
| 486 |
+
},
|
| 487 |
+
{
|
| 488 |
+
"epoch": 0.21,
|
| 489 |
+
"learning_rate": 4.804131442512309e-06,
|
| 490 |
+
"logits/chosen": 0.15109729766845703,
|
| 491 |
+
"logits/rejected": 0.17160046100616455,
|
| 492 |
+
"logps/chosen": -411.62109375,
|
| 493 |
+
"logps/rejected": -422.8251953125,
|
| 494 |
+
"loss": 0.1035,
|
| 495 |
+
"rewards/accuracies": 0.625,
|
| 496 |
+
"rewards/chosen": -0.08082294464111328,
|
| 497 |
+
"rewards/margins": 0.03992414474487305,
|
| 498 |
+
"rewards/rejected": -0.12074708938598633,
|
| 499 |
+
"step": 340
|
| 500 |
+
},
|
| 501 |
+
{
|
| 502 |
+
"epoch": 0.22,
|
| 503 |
+
"learning_rate": 4.78221901815689e-06,
|
| 504 |
+
"logits/chosen": 0.11030396074056625,
|
| 505 |
+
"logits/rejected": 0.10951008647680283,
|
| 506 |
+
"logps/chosen": -354.53839111328125,
|
| 507 |
+
"logps/rejected": -379.4768371582031,
|
| 508 |
+
"loss": 0.0944,
|
| 509 |
+
"rewards/accuracies": 0.53125,
|
| 510 |
+
"rewards/chosen": -0.07093557715415955,
|
| 511 |
+
"rewards/margins": 0.02872433327138424,
|
| 512 |
+
"rewards/rejected": -0.09965990483760834,
|
| 513 |
+
"step": 350
|
| 514 |
+
},
|
| 515 |
+
{
|
| 516 |
+
"epoch": 0.23,
|
| 517 |
+
"learning_rate": 4.7592005009687836e-06,
|
| 518 |
+
"logits/chosen": 0.12368909269571304,
|
| 519 |
+
"logits/rejected": 0.1551579087972641,
|
| 520 |
+
"logps/chosen": -422.1844787597656,
|
| 521 |
+
"logps/rejected": -442.5235290527344,
|
| 522 |
+
"loss": 0.0942,
|
| 523 |
+
"rewards/accuracies": 0.606249988079071,
|
| 524 |
+
"rewards/chosen": -0.09620015323162079,
|
| 525 |
+
"rewards/margins": 0.03423624113202095,
|
| 526 |
+
"rewards/rejected": -0.13043639063835144,
|
| 527 |
+
"step": 360
|
| 528 |
+
},
|
| 529 |
+
{
|
| 530 |
+
"epoch": 0.23,
|
| 531 |
+
"learning_rate": 4.735087047027525e-06,
|
| 532 |
+
"logits/chosen": 0.09877191483974457,
|
| 533 |
+
"logits/rejected": 0.1075328141450882,
|
| 534 |
+
"logps/chosen": -345.7369689941406,
|
| 535 |
+
"logps/rejected": -428.72320556640625,
|
| 536 |
+
"loss": 0.1058,
|
| 537 |
+
"rewards/accuracies": 0.5375000238418579,
|
| 538 |
+
"rewards/chosen": -0.08192293345928192,
|
| 539 |
+
"rewards/margins": 0.041438352316617966,
|
| 540 |
+
"rewards/rejected": -0.12336127460002899,
|
| 541 |
+
"step": 370
|
| 542 |
+
},
|
| 543 |
+
{
|
| 544 |
+
"epoch": 0.24,
|
| 545 |
+
"learning_rate": 4.709890343081127e-06,
|
| 546 |
+
"logits/chosen": 0.11233127117156982,
|
| 547 |
+
"logits/rejected": 0.09415660798549652,
|
| 548 |
+
"logps/chosen": -375.85552978515625,
|
| 549 |
+
"logps/rejected": -400.69561767578125,
|
| 550 |
+
"loss": 0.0912,
|
| 551 |
+
"rewards/accuracies": 0.550000011920929,
|
| 552 |
+
"rewards/chosen": -0.07030915468931198,
|
| 553 |
+
"rewards/margins": 0.028951376676559448,
|
| 554 |
+
"rewards/rejected": -0.09926052391529083,
|
| 555 |
+
"step": 380
|
| 556 |
+
},
|
| 557 |
+
{
|
| 558 |
+
"epoch": 0.25,
|
| 559 |
+
"learning_rate": 4.6836226008820215e-06,
|
| 560 |
+
"logits/chosen": 0.16574355959892273,
|
| 561 |
+
"logits/rejected": 0.1816074252128601,
|
| 562 |
+
"logps/chosen": -416.6439514160156,
|
| 563 |
+
"logps/rejected": -431.9356994628906,
|
| 564 |
+
"loss": 0.0968,
|
| 565 |
+
"rewards/accuracies": 0.65625,
|
| 566 |
+
"rewards/chosen": -0.06910283863544464,
|
| 567 |
+
"rewards/margins": 0.021879449486732483,
|
| 568 |
+
"rewards/rejected": -0.09098228812217712,
|
| 569 |
+
"step": 390
|
| 570 |
+
},
|
| 571 |
+
{
|
| 572 |
+
"epoch": 0.25,
|
| 573 |
+
"learning_rate": 4.656296551268547e-06,
|
| 574 |
+
"logits/chosen": 0.15201489627361298,
|
| 575 |
+
"logits/rejected": 0.18494170904159546,
|
| 576 |
+
"logps/chosen": -382.65350341796875,
|
| 577 |
+
"logps/rejected": -439.33502197265625,
|
| 578 |
+
"loss": 0.0938,
|
| 579 |
+
"rewards/accuracies": 0.612500011920929,
|
| 580 |
+
"rewards/chosen": -0.0796704962849617,
|
| 581 |
+
"rewards/margins": 0.04547972232103348,
|
| 582 |
+
"rewards/rejected": -0.12515023350715637,
|
| 583 |
+
"step": 400
|
| 584 |
+
},
|
| 585 |
+
{
|
| 586 |
+
"epoch": 0.26,
|
| 587 |
+
"learning_rate": 4.627925437994868e-06,
|
| 588 |
+
"logits/chosen": 0.13324889540672302,
|
| 589 |
+
"logits/rejected": 0.15600983798503876,
|
| 590 |
+
"logps/chosen": -342.90924072265625,
|
| 591 |
+
"logps/rejected": -418.2884216308594,
|
| 592 |
+
"loss": 0.0905,
|
| 593 |
+
"rewards/accuracies": 0.574999988079071,
|
| 594 |
+
"rewards/chosen": -0.08169208467006683,
|
| 595 |
+
"rewards/margins": 0.04994327947497368,
|
| 596 |
+
"rewards/rejected": -0.1316353678703308,
|
| 597 |
+
"step": 410
|
| 598 |
+
},
|
| 599 |
+
{
|
| 600 |
+
"epoch": 0.26,
|
| 601 |
+
"learning_rate": 4.598523011312296e-06,
|
| 602 |
+
"logits/chosen": 0.15302890539169312,
|
| 603 |
+
"logits/rejected": 0.15263435244560242,
|
| 604 |
+
"logps/chosen": -399.0650939941406,
|
| 605 |
+
"logps/rejected": -442.39080810546875,
|
| 606 |
+
"loss": 0.0898,
|
| 607 |
+
"rewards/accuracies": 0.550000011920929,
|
| 608 |
+
"rewards/chosen": -0.08143939077854156,
|
| 609 |
+
"rewards/margins": 0.03956509754061699,
|
| 610 |
+
"rewards/rejected": -0.12100447714328766,
|
| 611 |
+
"step": 420
|
| 612 |
+
},
|
| 613 |
+
{
|
| 614 |
+
"epoch": 0.27,
|
| 615 |
+
"learning_rate": 4.568103521305137e-06,
|
| 616 |
+
"logits/chosen": 0.18960033357143402,
|
| 617 |
+
"logits/rejected": 0.15051600337028503,
|
| 618 |
+
"logps/chosen": -382.1358337402344,
|
| 619 |
+
"logps/rejected": -444.8484802246094,
|
| 620 |
+
"loss": 0.0959,
|
| 621 |
+
"rewards/accuracies": 0.512499988079071,
|
| 622 |
+
"rewards/chosen": -0.07425881922245026,
|
| 623 |
+
"rewards/margins": 0.04583179950714111,
|
| 624 |
+
"rewards/rejected": -0.12009060382843018,
|
| 625 |
+
"step": 430
|
| 626 |
+
},
|
| 627 |
+
{
|
| 628 |
+
"epoch": 0.28,
|
| 629 |
+
"learning_rate": 4.536681710984296e-06,
|
| 630 |
+
"logits/chosen": 0.14668092131614685,
|
| 631 |
+
"logits/rejected": 0.20314042270183563,
|
| 632 |
+
"logps/chosen": -405.0791015625,
|
| 633 |
+
"logps/rejected": -454.7623596191406,
|
| 634 |
+
"loss": 0.0927,
|
| 635 |
+
"rewards/accuracies": 0.612500011920929,
|
| 636 |
+
"rewards/chosen": -0.07047592103481293,
|
| 637 |
+
"rewards/margins": 0.05656527355313301,
|
| 638 |
+
"rewards/rejected": -0.12704120576381683,
|
| 639 |
+
"step": 440
|
| 640 |
+
},
|
| 641 |
+
{
|
| 642 |
+
"epoch": 0.28,
|
| 643 |
+
"learning_rate": 4.5042728091419764e-06,
|
| 644 |
+
"logits/chosen": 0.14518819749355316,
|
| 645 |
+
"logits/rejected": 0.15771886706352234,
|
| 646 |
+
"logps/chosen": -351.4585266113281,
|
| 647 |
+
"logps/rejected": -418.5862731933594,
|
| 648 |
+
"loss": 0.0903,
|
| 649 |
+
"rewards/accuracies": 0.6000000238418579,
|
| 650 |
+
"rewards/chosen": -0.06795553117990494,
|
| 651 |
+
"rewards/margins": 0.03801413252949715,
|
| 652 |
+
"rewards/rejected": -0.10596966743469238,
|
| 653 |
+
"step": 450
|
| 654 |
+
},
|
| 655 |
+
{
|
| 656 |
+
"epoch": 0.29,
|
| 657 |
+
"learning_rate": 4.470892522970937e-06,
|
| 658 |
+
"logits/chosen": 0.12991555035114288,
|
| 659 |
+
"logits/rejected": 0.13929405808448792,
|
| 660 |
+
"logps/chosen": -369.97943115234375,
|
| 661 |
+
"logps/rejected": -418.1014099121094,
|
| 662 |
+
"loss": 0.0993,
|
| 663 |
+
"rewards/accuracies": 0.574999988079071,
|
| 664 |
+
"rewards/chosen": -0.06941680610179901,
|
| 665 |
+
"rewards/margins": 0.03608622029423714,
|
| 666 |
+
"rewards/rejected": -0.10550303757190704,
|
| 667 |
+
"step": 460
|
| 668 |
+
},
|
| 669 |
+
{
|
| 670 |
+
"epoch": 0.3,
|
| 671 |
+
"learning_rate": 4.436557030451907e-06,
|
| 672 |
+
"logits/chosen": 0.13183408975601196,
|
| 673 |
+
"logits/rejected": 0.14138123393058777,
|
| 674 |
+
"logps/chosen": -386.870849609375,
|
| 675 |
+
"logps/rejected": -483.94854736328125,
|
| 676 |
+
"loss": 0.1037,
|
| 677 |
+
"rewards/accuracies": 0.6312500238418579,
|
| 678 |
+
"rewards/chosen": -0.08128295838832855,
|
| 679 |
+
"rewards/margins": 0.05732026696205139,
|
| 680 |
+
"rewards/rejected": -0.13860321044921875,
|
| 681 |
+
"step": 470
|
| 682 |
+
},
|
| 683 |
+
{
|
| 684 |
+
"epoch": 0.3,
|
| 685 |
+
"learning_rate": 4.401282972512805e-06,
|
| 686 |
+
"logits/chosen": 0.09972018748521805,
|
| 687 |
+
"logits/rejected": 0.10249191522598267,
|
| 688 |
+
"logps/chosen": -423.345703125,
|
| 689 |
+
"logps/rejected": -447.601318359375,
|
| 690 |
+
"loss": 0.1098,
|
| 691 |
+
"rewards/accuracies": 0.574999988079071,
|
| 692 |
+
"rewards/chosen": -0.09919498860836029,
|
| 693 |
+
"rewards/margins": 0.02907257154583931,
|
| 694 |
+
"rewards/rejected": -0.1282675564289093,
|
| 695 |
+
"step": 480
|
| 696 |
+
},
|
| 697 |
+
{
|
| 698 |
+
"epoch": 0.31,
|
| 699 |
+
"learning_rate": 4.365087444963611e-06,
|
| 700 |
+
"logits/chosen": 0.08806703239679337,
|
| 701 |
+
"logits/rejected": 0.1713884323835373,
|
| 702 |
+
"logps/chosen": -467.56182861328125,
|
| 703 |
+
"logps/rejected": -477.89581298828125,
|
| 704 |
+
"loss": 0.0904,
|
| 705 |
+
"rewards/accuracies": 0.612500011920929,
|
| 706 |
+
"rewards/chosen": -0.09871919453144073,
|
| 707 |
+
"rewards/margins": 0.04472241550683975,
|
| 708 |
+
"rewards/rejected": -0.14344163239002228,
|
| 709 |
+
"step": 490
|
| 710 |
+
},
|
| 711 |
+
{
|
| 712 |
+
"epoch": 0.32,
|
| 713 |
+
"learning_rate": 4.327987990210763e-06,
|
| 714 |
+
"logits/chosen": 0.126537024974823,
|
| 715 |
+
"logits/rejected": 0.15067018568515778,
|
| 716 |
+
"logps/chosen": -361.6537170410156,
|
| 717 |
+
"logps/rejected": -433.5013122558594,
|
| 718 |
+
"loss": 0.0984,
|
| 719 |
+
"rewards/accuracies": 0.668749988079071,
|
| 720 |
+
"rewards/chosen": -0.08296778053045273,
|
| 721 |
+
"rewards/margins": 0.053381673991680145,
|
| 722 |
+
"rewards/rejected": -0.13634945452213287,
|
| 723 |
+
"step": 500
|
| 724 |
+
},
|
| 725 |
+
{
|
| 726 |
+
"epoch": 0.32,
|
| 727 |
+
"learning_rate": 4.290002588755112e-06,
|
| 728 |
+
"logits/chosen": 0.14435867965221405,
|
| 729 |
+
"logits/rejected": 0.18025091290473938,
|
| 730 |
+
"logps/chosen": -383.94110107421875,
|
| 731 |
+
"logps/rejected": -440.88446044921875,
|
| 732 |
+
"loss": 0.0917,
|
| 733 |
+
"rewards/accuracies": 0.6187499761581421,
|
| 734 |
+
"rewards/chosen": -0.07438959181308746,
|
| 735 |
+
"rewards/margins": 0.035205405205488205,
|
| 736 |
+
"rewards/rejected": -0.10959500074386597,
|
| 737 |
+
"step": 510
|
| 738 |
+
},
|
| 739 |
+
{
|
| 740 |
+
"epoch": 0.33,
|
| 741 |
+
"learning_rate": 4.251149650477553e-06,
|
| 742 |
+
"logits/chosen": 0.11688809096813202,
|
| 743 |
+
"logits/rejected": 0.1297851949930191,
|
| 744 |
+
"logps/chosen": -420.3448181152344,
|
| 745 |
+
"logps/rejected": -450.2854919433594,
|
| 746 |
+
"loss": 0.1182,
|
| 747 |
+
"rewards/accuracies": 0.550000011920929,
|
| 748 |
+
"rewards/chosen": -0.10054901987314224,
|
| 749 |
+
"rewards/margins": 0.025883013382554054,
|
| 750 |
+
"rewards/rejected": -0.12643203139305115,
|
| 751 |
+
"step": 520
|
| 752 |
+
},
|
| 753 |
+
{
|
| 754 |
+
"epoch": 0.33,
|
| 755 |
+
"learning_rate": 4.211448005716547e-06,
|
| 756 |
+
"logits/chosen": 0.140035018324852,
|
| 757 |
+
"logits/rejected": 0.1768953651189804,
|
| 758 |
+
"logps/chosen": -340.6114196777344,
|
| 759 |
+
"logps/rejected": -433.100341796875,
|
| 760 |
+
"loss": 0.1061,
|
| 761 |
+
"rewards/accuracies": 0.625,
|
| 762 |
+
"rewards/chosen": -0.06748305261135101,
|
| 763 |
+
"rewards/margins": 0.05944541096687317,
|
| 764 |
+
"rewards/rejected": -0.12692847847938538,
|
| 765 |
+
"step": 530
|
| 766 |
+
},
|
| 767 |
+
{
|
| 768 |
+
"epoch": 0.34,
|
| 769 |
+
"learning_rate": 4.170916896141877e-06,
|
| 770 |
+
"logits/chosen": 0.11086130142211914,
|
| 771 |
+
"logits/rejected": 0.11766747385263443,
|
| 772 |
+
"logps/chosen": -409.5714111328125,
|
| 773 |
+
"logps/rejected": -471.7931213378906,
|
| 774 |
+
"loss": 0.0909,
|
| 775 |
+
"rewards/accuracies": 0.612500011920929,
|
| 776 |
+
"rewards/chosen": -0.08654967695474625,
|
| 777 |
+
"rewards/margins": 0.052199166268110275,
|
| 778 |
+
"rewards/rejected": -0.13874883949756622,
|
| 779 |
+
"step": 540
|
| 780 |
+
},
|
| 781 |
+
{
|
| 782 |
+
"epoch": 0.35,
|
| 783 |
+
"learning_rate": 4.129575965429036e-06,
|
| 784 |
+
"logits/chosen": 0.1578402817249298,
|
| 785 |
+
"logits/rejected": 0.12391265481710434,
|
| 786 |
+
"logps/chosen": -406.35784912109375,
|
| 787 |
+
"logps/rejected": -423.5259704589844,
|
| 788 |
+
"loss": 0.0988,
|
| 789 |
+
"rewards/accuracies": 0.606249988079071,
|
| 790 |
+
"rewards/chosen": -0.07460321485996246,
|
| 791 |
+
"rewards/margins": 0.04027136042714119,
|
| 792 |
+
"rewards/rejected": -0.11487456411123276,
|
| 793 |
+
"step": 550
|
| 794 |
+
},
|
| 795 |
+
{
|
| 796 |
+
"epoch": 0.35,
|
| 797 |
+
"learning_rate": 4.087445249738795e-06,
|
| 798 |
+
"logits/chosen": 0.13769268989562988,
|
| 799 |
+
"logits/rejected": 0.14535634219646454,
|
| 800 |
+
"logps/chosen": -402.810791015625,
|
| 801 |
+
"logps/rejected": -468.11578369140625,
|
| 802 |
+
"loss": 0.0885,
|
| 803 |
+
"rewards/accuracies": 0.6499999761581421,
|
| 804 |
+
"rewards/chosen": -0.07266028970479965,
|
| 805 |
+
"rewards/margins": 0.04542839899659157,
|
| 806 |
+
"rewards/rejected": -0.11808867752552032,
|
| 807 |
+
"step": 560
|
| 808 |
+
},
|
| 809 |
+
{
|
| 810 |
+
"epoch": 0.36,
|
| 811 |
+
"learning_rate": 4.044545168006535e-06,
|
| 812 |
+
"logits/chosen": 0.135701984167099,
|
| 813 |
+
"logits/rejected": 0.15048106014728546,
|
| 814 |
+
"logps/chosen": -367.40716552734375,
|
| 815 |
+
"logps/rejected": -425.750732421875,
|
| 816 |
+
"loss": 0.0933,
|
| 817 |
+
"rewards/accuracies": 0.612500011920929,
|
| 818 |
+
"rewards/chosen": -0.06163717433810234,
|
| 819 |
+
"rewards/margins": 0.0523550882935524,
|
| 820 |
+
"rewards/rejected": -0.11399225890636444,
|
| 821 |
+
"step": 570
|
| 822 |
+
},
|
| 823 |
+
{
|
| 824 |
+
"epoch": 0.37,
|
| 825 |
+
"learning_rate": 4.000896512046079e-06,
|
| 826 |
+
"logits/chosen": 0.06643296778202057,
|
| 827 |
+
"logits/rejected": 0.11304454505443573,
|
| 828 |
+
"logps/chosen": -382.0234069824219,
|
| 829 |
+
"logps/rejected": -408.7083435058594,
|
| 830 |
+
"loss": 0.0964,
|
| 831 |
+
"rewards/accuracies": 0.643750011920929,
|
| 832 |
+
"rewards/chosen": -0.08725063502788544,
|
| 833 |
+
"rewards/margins": 0.030442586168646812,
|
| 834 |
+
"rewards/rejected": -0.117693230509758,
|
| 835 |
+
"step": 580
|
| 836 |
+
},
|
| 837 |
+
{
|
| 838 |
+
"epoch": 0.37,
|
| 839 |
+
"learning_rate": 3.956520436472794e-06,
|
| 840 |
+
"logits/chosen": 0.09026265144348145,
|
| 841 |
+
"logits/rejected": 0.14468029141426086,
|
| 842 |
+
"logps/chosen": -425.26214599609375,
|
| 843 |
+
"logps/rejected": -457.64990234375,
|
| 844 |
+
"loss": 0.0923,
|
| 845 |
+
"rewards/accuracies": 0.675000011920929,
|
| 846 |
+
"rewards/chosen": -0.09045656025409698,
|
| 847 |
+
"rewards/margins": 0.050621867179870605,
|
| 848 |
+
"rewards/rejected": -0.1410784274339676,
|
| 849 |
+
"step": 590
|
| 850 |
+
},
|
| 851 |
+
{
|
| 852 |
+
"epoch": 0.38,
|
| 853 |
+
"learning_rate": 3.911438448450867e-06,
|
| 854 |
+
"logits/chosen": 0.09659644216299057,
|
| 855 |
+
"logits/rejected": 0.10519146919250488,
|
| 856 |
+
"logps/chosen": -409.36737060546875,
|
| 857 |
+
"logps/rejected": -426.3362731933594,
|
| 858 |
+
"loss": 0.0965,
|
| 859 |
+
"rewards/accuracies": 0.550000011920929,
|
| 860 |
+
"rewards/chosen": -0.09249825030565262,
|
| 861 |
+
"rewards/margins": 0.03923904523253441,
|
| 862 |
+
"rewards/rejected": -0.13173729181289673,
|
| 863 |
+
"step": 600
|
| 864 |
+
},
|
| 865 |
+
{
|
| 866 |
+
"epoch": 0.38,
|
| 867 |
+
"learning_rate": 3.865672397269712e-06,
|
| 868 |
+
"logits/chosen": 0.16088059544563293,
|
| 869 |
+
"logits/rejected": 0.15578779578208923,
|
| 870 |
+
"logps/chosen": -421.38134765625,
|
| 871 |
+
"logps/rejected": -467.1226501464844,
|
| 872 |
+
"loss": 0.103,
|
| 873 |
+
"rewards/accuracies": 0.6312500238418579,
|
| 874 |
+
"rewards/chosen": -0.09401367604732513,
|
| 875 |
+
"rewards/margins": 0.037443626672029495,
|
| 876 |
+
"rewards/rejected": -0.13145729899406433,
|
| 877 |
+
"step": 610
|
| 878 |
+
},
|
| 879 |
+
{
|
| 880 |
+
"epoch": 0.39,
|
| 881 |
+
"learning_rate": 3.819244463754566e-06,
|
| 882 |
+
"logits/chosen": 0.1309213638305664,
|
| 883 |
+
"logits/rejected": 0.183641254901886,
|
| 884 |
+
"logps/chosen": -372.6539611816406,
|
| 885 |
+
"logps/rejected": -422.17333984375,
|
| 886 |
+
"loss": 0.1079,
|
| 887 |
+
"rewards/accuracies": 0.5625,
|
| 888 |
+
"rewards/chosen": -0.07998184859752655,
|
| 889 |
+
"rewards/margins": 0.03546987846493721,
|
| 890 |
+
"rewards/rejected": -0.11545171588659286,
|
| 891 |
+
"step": 620
|
| 892 |
+
},
|
| 893 |
+
{
|
| 894 |
+
"epoch": 0.4,
|
| 895 |
+
"learning_rate": 3.7721771495164e-06,
|
| 896 |
+
"logits/chosen": 0.18562278151512146,
|
| 897 |
+
"logits/rejected": 0.17855581641197205,
|
| 898 |
+
"logps/chosen": -368.06634521484375,
|
| 899 |
+
"logps/rejected": -410.43157958984375,
|
| 900 |
+
"loss": 0.0984,
|
| 901 |
+
"rewards/accuracies": 0.581250011920929,
|
| 902 |
+
"rewards/chosen": -0.06547744572162628,
|
| 903 |
+
"rewards/margins": 0.037668634206056595,
|
| 904 |
+
"rewards/rejected": -0.10314607620239258,
|
| 905 |
+
"step": 630
|
| 906 |
+
},
|
| 907 |
+
{
|
| 908 |
+
"epoch": 0.4,
|
| 909 |
+
"learning_rate": 3.7244932660463628e-06,
|
| 910 |
+
"logits/chosen": 0.10567553341388702,
|
| 911 |
+
"logits/rejected": 0.09214256703853607,
|
| 912 |
+
"logps/chosen": -387.8129577636719,
|
| 913 |
+
"logps/rejected": -428.2767028808594,
|
| 914 |
+
"loss": 0.1038,
|
| 915 |
+
"rewards/accuracies": 0.606249988079071,
|
| 916 |
+
"rewards/chosen": -0.08541779965162277,
|
| 917 |
+
"rewards/margins": 0.04381578415632248,
|
| 918 |
+
"rewards/rejected": -0.12923358380794525,
|
| 919 |
+
"step": 640
|
| 920 |
+
},
|
| 921 |
+
{
|
| 922 |
+
"epoch": 0.41,
|
| 923 |
+
"learning_rate": 3.676215923660037e-06,
|
| 924 |
+
"logits/chosen": 0.1147516518831253,
|
| 925 |
+
"logits/rejected": 0.13854558765888214,
|
| 926 |
+
"logps/chosen": -388.88397216796875,
|
| 927 |
+
"logps/rejected": -402.66162109375,
|
| 928 |
+
"loss": 0.0979,
|
| 929 |
+
"rewards/accuracies": 0.5249999761581421,
|
| 930 |
+
"rewards/chosen": -0.08885400742292404,
|
| 931 |
+
"rewards/margins": 0.03468873351812363,
|
| 932 |
+
"rewards/rejected": -0.12354274094104767,
|
| 933 |
+
"step": 650
|
| 934 |
+
},
|
| 935 |
+
{
|
| 936 |
+
"epoch": 0.42,
|
| 937 |
+
"learning_rate": 3.6273685202968727e-06,
|
| 938 |
+
"logits/chosen": 0.14652499556541443,
|
| 939 |
+
"logits/rejected": 0.23127985000610352,
|
| 940 |
+
"logps/chosen": -427.195556640625,
|
| 941 |
+
"logps/rejected": -474.3085021972656,
|
| 942 |
+
"loss": 0.1012,
|
| 943 |
+
"rewards/accuracies": 0.6187499761581421,
|
| 944 |
+
"rewards/chosen": -0.10611389577388763,
|
| 945 |
+
"rewards/margins": 0.047405023127794266,
|
| 946 |
+
"rewards/rejected": -0.1535189151763916,
|
| 947 |
+
"step": 660
|
| 948 |
+
},
|
| 949 |
+
{
|
| 950 |
+
"epoch": 0.42,
|
| 951 |
+
"learning_rate": 3.577974730180212e-06,
|
| 952 |
+
"logits/chosen": 0.1295718103647232,
|
| 953 |
+
"logits/rejected": 0.18317390978336334,
|
| 954 |
+
"logps/chosen": -454.80938720703125,
|
| 955 |
+
"logps/rejected": -480.46490478515625,
|
| 956 |
+
"loss": 0.1008,
|
| 957 |
+
"rewards/accuracies": 0.668749988079071,
|
| 958 |
+
"rewards/chosen": -0.09846655279397964,
|
| 959 |
+
"rewards/margins": 0.038999587297439575,
|
| 960 |
+
"rewards/rejected": -0.13746614754199982,
|
| 961 |
+
"step": 670
|
| 962 |
+
},
|
| 963 |
+
{
|
| 964 |
+
"epoch": 0.43,
|
| 965 |
+
"learning_rate": 3.528058492343423e-06,
|
| 966 |
+
"logits/chosen": 0.11106806993484497,
|
| 967 |
+
"logits/rejected": 0.1717728078365326,
|
| 968 |
+
"logps/chosen": -415.18414306640625,
|
| 969 |
+
"logps/rejected": -438.2942810058594,
|
| 970 |
+
"loss": 0.1083,
|
| 971 |
+
"rewards/accuracies": 0.550000011920929,
|
| 972 |
+
"rewards/chosen": -0.1040610671043396,
|
| 973 |
+
"rewards/margins": 0.030150944367051125,
|
| 974 |
+
"rewards/rejected": -0.13421201705932617,
|
| 975 |
+
"step": 680
|
| 976 |
+
},
|
| 977 |
+
{
|
| 978 |
+
"epoch": 0.43,
|
| 979 |
+
"learning_rate": 3.477643999027683e-06,
|
| 980 |
+
"logits/chosen": 0.16643357276916504,
|
| 981 |
+
"logits/rejected": 0.18508140742778778,
|
| 982 |
+
"logps/chosen": -382.0606384277344,
|
| 983 |
+
"logps/rejected": -410.14691162109375,
|
| 984 |
+
"loss": 0.0969,
|
| 985 |
+
"rewards/accuracies": 0.550000011920929,
|
| 986 |
+
"rewards/chosen": -0.08891114592552185,
|
| 987 |
+
"rewards/margins": 0.03431970998644829,
|
| 988 |
+
"rewards/rejected": -0.12323085963726044,
|
| 989 |
+
"step": 690
|
| 990 |
+
},
|
| 991 |
+
{
|
| 992 |
+
"epoch": 0.44,
|
| 993 |
+
"learning_rate": 3.4267556839570467e-06,
|
| 994 |
+
"logits/chosen": 0.11085295677185059,
|
| 995 |
+
"logits/rejected": 0.11592169851064682,
|
| 996 |
+
"logps/chosen": -380.14892578125,
|
| 997 |
+
"logps/rejected": -436.01422119140625,
|
| 998 |
+
"loss": 0.0928,
|
| 999 |
+
"rewards/accuracies": 0.606249988079071,
|
| 1000 |
+
"rewards/chosen": -0.08309037983417511,
|
| 1001 |
+
"rewards/margins": 0.050611067563295364,
|
| 1002 |
+
"rewards/rejected": -0.13370142877101898,
|
| 1003 |
+
"step": 700
|
| 1004 |
+
},
|
| 1005 |
+
{
|
| 1006 |
+
"epoch": 0.45,
|
| 1007 |
+
"learning_rate": 3.375418210496473e-06,
|
| 1008 |
+
"logits/chosen": 0.14460065960884094,
|
| 1009 |
+
"logits/rejected": 0.17223772406578064,
|
| 1010 |
+
"logps/chosen": -410.51422119140625,
|
| 1011 |
+
"logps/rejected": -431.4485778808594,
|
| 1012 |
+
"loss": 0.1039,
|
| 1013 |
+
"rewards/accuracies": 0.5874999761581421,
|
| 1014 |
+
"rewards/chosen": -0.08633999526500702,
|
| 1015 |
+
"rewards/margins": 0.03471770137548447,
|
| 1016 |
+
"rewards/rejected": -0.12105770409107208,
|
| 1017 |
+
"step": 710
|
| 1018 |
+
},
|
| 1019 |
+
{
|
| 1020 |
+
"epoch": 0.45,
|
| 1021 |
+
"learning_rate": 3.3236564596985677e-06,
|
| 1022 |
+
"logits/chosen": 0.11466125398874283,
|
| 1023 |
+
"logits/rejected": 0.14213159680366516,
|
| 1024 |
+
"logps/chosen": -429.1424865722656,
|
| 1025 |
+
"logps/rejected": -489.776123046875,
|
| 1026 |
+
"loss": 0.083,
|
| 1027 |
+
"rewards/accuracies": 0.6499999761581421,
|
| 1028 |
+
"rewards/chosen": -0.08991511166095734,
|
| 1029 |
+
"rewards/margins": 0.04996410012245178,
|
| 1030 |
+
"rewards/rejected": -0.13987921178340912,
|
| 1031 |
+
"step": 720
|
| 1032 |
+
},
|
| 1033 |
+
{
|
| 1034 |
+
"epoch": 0.46,
|
| 1035 |
+
"learning_rate": 3.2714955182448044e-06,
|
| 1036 |
+
"logits/chosen": 0.06632956117391586,
|
| 1037 |
+
"logits/rejected": 0.10460522025823593,
|
| 1038 |
+
"logps/chosen": -355.56988525390625,
|
| 1039 |
+
"logps/rejected": -453.8470764160156,
|
| 1040 |
+
"loss": 0.1004,
|
| 1041 |
+
"rewards/accuracies": 0.637499988079071,
|
| 1042 |
+
"rewards/chosen": -0.07422816753387451,
|
| 1043 |
+
"rewards/margins": 0.059829723089933395,
|
| 1044 |
+
"rewards/rejected": -0.1340578943490982,
|
| 1045 |
+
"step": 730
|
| 1046 |
+
},
|
| 1047 |
+
{
|
| 1048 |
+
"epoch": 0.47,
|
| 1049 |
+
"learning_rate": 3.2189606662871017e-06,
|
| 1050 |
+
"logits/chosen": 0.1409912258386612,
|
| 1051 |
+
"logits/rejected": 0.14539703726768494,
|
| 1052 |
+
"logps/chosen": -390.64251708984375,
|
| 1053 |
+
"logps/rejected": -435.5365295410156,
|
| 1054 |
+
"loss": 0.0906,
|
| 1055 |
+
"rewards/accuracies": 0.6187499761581421,
|
| 1056 |
+
"rewards/chosen": -0.07141707837581635,
|
| 1057 |
+
"rewards/margins": 0.03945548087358475,
|
| 1058 |
+
"rewards/rejected": -0.1108725517988205,
|
| 1059 |
+
"step": 740
|
| 1060 |
+
},
|
| 1061 |
+
{
|
| 1062 |
+
"epoch": 0.47,
|
| 1063 |
+
"learning_rate": 3.1660773651956223e-06,
|
| 1064 |
+
"logits/chosen": 0.12456997483968735,
|
| 1065 |
+
"logits/rejected": 0.1372366100549698,
|
| 1066 |
+
"logps/chosen": -355.9635314941406,
|
| 1067 |
+
"logps/rejected": -383.43853759765625,
|
| 1068 |
+
"loss": 0.1004,
|
| 1069 |
+
"rewards/accuracies": 0.550000011920929,
|
| 1070 |
+
"rewards/chosen": -0.07638130336999893,
|
| 1071 |
+
"rewards/margins": 0.037354812026023865,
|
| 1072 |
+
"rewards/rejected": -0.1137361153960228,
|
| 1073 |
+
"step": 750
|
| 1074 |
+
},
|
| 1075 |
+
{
|
| 1076 |
+
"epoch": 0.48,
|
| 1077 |
+
"learning_rate": 3.1128712452187535e-06,
|
| 1078 |
+
"logits/chosen": 0.09267710149288177,
|
| 1079 |
+
"logits/rejected": 0.07987795025110245,
|
| 1080 |
+
"logps/chosen": -402.697509765625,
|
| 1081 |
+
"logps/rejected": -450.74786376953125,
|
| 1082 |
+
"loss": 0.0967,
|
| 1083 |
+
"rewards/accuracies": 0.59375,
|
| 1084 |
+
"rewards/chosen": -0.08393266797065735,
|
| 1085 |
+
"rewards/margins": 0.04311661794781685,
|
| 1086 |
+
"rewards/rejected": -0.1270492970943451,
|
| 1087 |
+
"step": 760
|
| 1088 |
+
},
|
| 1089 |
+
{
|
| 1090 |
+
"epoch": 0.49,
|
| 1091 |
+
"learning_rate": 3.0593680930612295e-06,
|
| 1092 |
+
"logits/chosen": 0.0781303271651268,
|
| 1093 |
+
"logits/rejected": 0.16469840705394745,
|
| 1094 |
+
"logps/chosen": -437.64013671875,
|
| 1095 |
+
"logps/rejected": -474.09661865234375,
|
| 1096 |
+
"loss": 0.1021,
|
| 1097 |
+
"rewards/accuracies": 0.59375,
|
| 1098 |
+
"rewards/chosen": -0.10180965811014175,
|
| 1099 |
+
"rewards/margins": 0.03583382070064545,
|
| 1100 |
+
"rewards/rejected": -0.1376434713602066,
|
| 1101 |
+
"step": 770
|
| 1102 |
+
},
|
| 1103 |
+
{
|
| 1104 |
+
"epoch": 0.49,
|
| 1105 |
+
"learning_rate": 3.005593839386434e-06,
|
| 1106 |
+
"logits/chosen": 0.10138174146413803,
|
| 1107 |
+
"logits/rejected": 0.1595858335494995,
|
| 1108 |
+
"logps/chosen": -466.75811767578125,
|
| 1109 |
+
"logps/rejected": -494.52435302734375,
|
| 1110 |
+
"loss": 0.0913,
|
| 1111 |
+
"rewards/accuracies": 0.6187499761581421,
|
| 1112 |
+
"rewards/chosen": -0.09192174673080444,
|
| 1113 |
+
"rewards/margins": 0.048992156982421875,
|
| 1114 |
+
"rewards/rejected": -0.14091390371322632,
|
| 1115 |
+
"step": 780
|
| 1116 |
+
},
|
| 1117 |
+
{
|
| 1118 |
+
"epoch": 0.5,
|
| 1119 |
+
"learning_rate": 2.951574546248928e-06,
|
| 1120 |
+
"logits/chosen": 0.09503944218158722,
|
| 1121 |
+
"logits/rejected": 0.1364067792892456,
|
| 1122 |
+
"logps/chosen": -458.38775634765625,
|
| 1123 |
+
"logps/rejected": -459.333251953125,
|
| 1124 |
+
"loss": 0.0985,
|
| 1125 |
+
"rewards/accuracies": 0.5625,
|
| 1126 |
+
"rewards/chosen": -0.09930691868066788,
|
| 1127 |
+
"rewards/margins": 0.029138978570699692,
|
| 1128 |
+
"rewards/rejected": -0.12844590842723846,
|
| 1129 |
+
"step": 790
|
| 1130 |
+
},
|
| 1131 |
+
{
|
| 1132 |
+
"epoch": 0.5,
|
| 1133 |
+
"learning_rate": 2.897336394463301e-06,
|
| 1134 |
+
"logits/chosen": 0.10846900939941406,
|
| 1135 |
+
"logits/rejected": 0.09937886893749237,
|
| 1136 |
+
"logps/chosen": -424.815673828125,
|
| 1137 |
+
"logps/rejected": -459.7933044433594,
|
| 1138 |
+
"loss": 0.0941,
|
| 1139 |
+
"rewards/accuracies": 0.6625000238418579,
|
| 1140 |
+
"rewards/chosen": -0.08887843042612076,
|
| 1141 |
+
"rewards/margins": 0.05371922254562378,
|
| 1142 |
+
"rewards/rejected": -0.14259764552116394,
|
| 1143 |
+
"step": 800
|
| 1144 |
+
},
|
| 1145 |
+
{
|
| 1146 |
+
"epoch": 0.51,
|
| 1147 |
+
"learning_rate": 2.842905670915462e-06,
|
| 1148 |
+
"logits/chosen": 0.1058458462357521,
|
| 1149 |
+
"logits/rejected": 0.0782022550702095,
|
| 1150 |
+
"logps/chosen": -434.9790954589844,
|
| 1151 |
+
"logps/rejected": -469.34454345703125,
|
| 1152 |
+
"loss": 0.0967,
|
| 1153 |
+
"rewards/accuracies": 0.6187499761581421,
|
| 1154 |
+
"rewards/chosen": -0.09677759557962418,
|
| 1155 |
+
"rewards/margins": 0.04304363578557968,
|
| 1156 |
+
"rewards/rejected": -0.13982121646404266,
|
| 1157 |
+
"step": 810
|
| 1158 |
+
},
|
| 1159 |
+
{
|
| 1160 |
+
"epoch": 0.52,
|
| 1161 |
+
"learning_rate": 2.788308755822522e-06,
|
| 1162 |
+
"logits/chosen": 0.0967346802353859,
|
| 1163 |
+
"logits/rejected": 0.10881376266479492,
|
| 1164 |
+
"logps/chosen": -369.01995849609375,
|
| 1165 |
+
"logps/rejected": -449.07623291015625,
|
| 1166 |
+
"loss": 0.0926,
|
| 1167 |
+
"rewards/accuracies": 0.6312500238418579,
|
| 1168 |
+
"rewards/chosen": -0.09960204362869263,
|
| 1169 |
+
"rewards/margins": 0.04991160333156586,
|
| 1170 |
+
"rewards/rejected": -0.1495136320590973,
|
| 1171 |
+
"step": 820
|
| 1172 |
+
},
|
| 1173 |
+
{
|
| 1174 |
+
"epoch": 0.52,
|
| 1175 |
+
"learning_rate": 2.733572109947444e-06,
|
| 1176 |
+
"logits/chosen": 0.10651399195194244,
|
| 1177 |
+
"logits/rejected": 0.128664031624794,
|
| 1178 |
+
"logps/chosen": -466.0299377441406,
|
| 1179 |
+
"logps/rejected": -499.25762939453125,
|
| 1180 |
+
"loss": 0.0972,
|
| 1181 |
+
"rewards/accuracies": 0.6000000238418579,
|
| 1182 |
+
"rewards/chosen": -0.1191006675362587,
|
| 1183 |
+
"rewards/margins": 0.044090814888477325,
|
| 1184 |
+
"rewards/rejected": -0.16319148242473602,
|
| 1185 |
+
"step": 830
|
| 1186 |
+
},
|
| 1187 |
+
{
|
| 1188 |
+
"epoch": 0.53,
|
| 1189 |
+
"learning_rate": 2.6787222617746583e-06,
|
| 1190 |
+
"logits/chosen": 0.12026312202215195,
|
| 1191 |
+
"logits/rejected": 0.1642097681760788,
|
| 1192 |
+
"logps/chosen": -380.1086120605469,
|
| 1193 |
+
"logps/rejected": -439.4449157714844,
|
| 1194 |
+
"loss": 0.0884,
|
| 1195 |
+
"rewards/accuracies": 0.6000000238418579,
|
| 1196 |
+
"rewards/chosen": -0.09910561889410019,
|
| 1197 |
+
"rewards/margins": 0.04509486258029938,
|
| 1198 |
+
"rewards/rejected": -0.14420048892498016,
|
| 1199 |
+
"step": 840
|
| 1200 |
+
},
|
| 1201 |
+
{
|
| 1202 |
+
"epoch": 0.54,
|
| 1203 |
+
"learning_rate": 2.62378579465285e-06,
|
| 1204 |
+
"logits/chosen": 0.12080010026693344,
|
| 1205 |
+
"logits/rejected": 0.15950220823287964,
|
| 1206 |
+
"logps/chosen": -441.6851501464844,
|
| 1207 |
+
"logps/rejected": -498.75152587890625,
|
| 1208 |
+
"loss": 0.0919,
|
| 1209 |
+
"rewards/accuracies": 0.606249988079071,
|
| 1210 |
+
"rewards/chosen": -0.09932798892259598,
|
| 1211 |
+
"rewards/margins": 0.05052763968706131,
|
| 1212 |
+
"rewards/rejected": -0.14985564351081848,
|
| 1213 |
+
"step": 850
|
| 1214 |
+
},
|
| 1215 |
+
{
|
| 1216 |
+
"epoch": 0.54,
|
| 1217 |
+
"learning_rate": 2.568789333911165e-06,
|
| 1218 |
+
"logits/chosen": 0.1327330321073532,
|
| 1219 |
+
"logits/rejected": 0.14310786128044128,
|
| 1220 |
+
"logps/chosen": -375.2086486816406,
|
| 1221 |
+
"logps/rejected": -488.7005310058594,
|
| 1222 |
+
"loss": 0.0957,
|
| 1223 |
+
"rewards/accuracies": 0.637499988079071,
|
| 1224 |
+
"rewards/chosen": -0.08790336549282074,
|
| 1225 |
+
"rewards/margins": 0.0683864951133728,
|
| 1226 |
+
"rewards/rejected": -0.15628986060619354,
|
| 1227 |
+
"step": 860
|
| 1228 |
+
},
|
| 1229 |
+
{
|
| 1230 |
+
"epoch": 0.55,
|
| 1231 |
+
"learning_rate": 2.5137595339550564e-06,
|
| 1232 |
+
"logits/chosen": 0.10466363281011581,
|
| 1233 |
+
"logits/rejected": 0.12011264264583588,
|
| 1234 |
+
"logps/chosen": -457.970458984375,
|
| 1235 |
+
"logps/rejected": -444.8792419433594,
|
| 1236 |
+
"loss": 0.0943,
|
| 1237 |
+
"rewards/accuracies": 0.512499988079071,
|
| 1238 |
+
"rewards/chosen": -0.10931634902954102,
|
| 1239 |
+
"rewards/margins": 0.027462299913167953,
|
| 1240 |
+
"rewards/rejected": -0.13677865266799927,
|
| 1241 |
+
"step": 870
|
| 1242 |
+
},
|
| 1243 |
+
{
|
| 1244 |
+
"epoch": 0.55,
|
| 1245 |
+
"learning_rate": 2.4587230653480577e-06,
|
| 1246 |
+
"logits/chosen": 0.07321979105472565,
|
| 1247 |
+
"logits/rejected": 0.09171778708696365,
|
| 1248 |
+
"logps/chosen": -435.8045349121094,
|
| 1249 |
+
"logps/rejected": -464.2123107910156,
|
| 1250 |
+
"loss": 0.0896,
|
| 1251 |
+
"rewards/accuracies": 0.6625000238418579,
|
| 1252 |
+
"rewards/chosen": -0.11067759990692139,
|
| 1253 |
+
"rewards/margins": 0.045581571757793427,
|
| 1254 |
+
"rewards/rejected": -0.15625916421413422,
|
| 1255 |
+
"step": 880
|
| 1256 |
+
},
|
| 1257 |
+
{
|
| 1258 |
+
"epoch": 0.56,
|
| 1259 |
+
"learning_rate": 2.403706601885705e-06,
|
| 1260 |
+
"logits/chosen": 0.12501958012580872,
|
| 1261 |
+
"logits/rejected": 0.10027003288269043,
|
| 1262 |
+
"logps/chosen": -439.4681701660156,
|
| 1263 |
+
"logps/rejected": -450.8436584472656,
|
| 1264 |
+
"loss": 0.0988,
|
| 1265 |
+
"rewards/accuracies": 0.5874999761581421,
|
| 1266 |
+
"rewards/chosen": -0.09043081104755402,
|
| 1267 |
+
"rewards/margins": 0.03802182897925377,
|
| 1268 |
+
"rewards/rejected": -0.12845264375209808,
|
| 1269 |
+
"step": 890
|
| 1270 |
+
},
|
| 1271 |
+
{
|
| 1272 |
+
"epoch": 0.57,
|
| 1273 |
+
"learning_rate": 2.3487368076679095e-06,
|
| 1274 |
+
"logits/chosen": 0.03233950585126877,
|
| 1275 |
+
"logits/rejected": 0.10005395114421844,
|
| 1276 |
+
"logps/chosen": -423.36163330078125,
|
| 1277 |
+
"logps/rejected": -499.0625,
|
| 1278 |
+
"loss": 0.0823,
|
| 1279 |
+
"rewards/accuracies": 0.6187499761581421,
|
| 1280 |
+
"rewards/chosen": -0.10427310317754745,
|
| 1281 |
+
"rewards/margins": 0.07294373214244843,
|
| 1282 |
+
"rewards/rejected": -0.17721685767173767,
|
| 1283 |
+
"step": 900
|
| 1284 |
+
},
|
| 1285 |
+
{
|
| 1286 |
+
"epoch": 0.57,
|
| 1287 |
+
"learning_rate": 2.2938403241760153e-06,
|
| 1288 |
+
"logits/chosen": 0.046649765223264694,
|
| 1289 |
+
"logits/rejected": 0.06434310227632523,
|
| 1290 |
+
"logps/chosen": -385.28753662109375,
|
| 1291 |
+
"logps/rejected": -428.0790100097656,
|
| 1292 |
+
"loss": 0.0921,
|
| 1293 |
+
"rewards/accuracies": 0.6625000238418579,
|
| 1294 |
+
"rewards/chosen": -0.097981296479702,
|
| 1295 |
+
"rewards/margins": 0.046177938580513,
|
| 1296 |
+
"rewards/rejected": -0.1441592425107956,
|
| 1297 |
+
"step": 910
|
| 1298 |
+
},
|
| 1299 |
+
{
|
| 1300 |
+
"epoch": 0.58,
|
| 1301 |
+
"learning_rate": 2.2390437573608313e-06,
|
| 1302 |
+
"logits/chosen": 0.04348721355199814,
|
| 1303 |
+
"logits/rejected": 0.09973670542240143,
|
| 1304 |
+
"logps/chosen": -419.55841064453125,
|
| 1305 |
+
"logps/rejected": -455.224853515625,
|
| 1306 |
+
"loss": 0.0924,
|
| 1307 |
+
"rewards/accuracies": 0.5562499761581421,
|
| 1308 |
+
"rewards/chosen": -0.11863915622234344,
|
| 1309 |
+
"rewards/margins": 0.03215927258133888,
|
| 1310 |
+
"rewards/rejected": -0.15079843997955322,
|
| 1311 |
+
"step": 920
|
| 1312 |
+
},
|
| 1313 |
+
{
|
| 1314 |
+
"epoch": 0.59,
|
| 1315 |
+
"learning_rate": 2.1843736647478733e-06,
|
| 1316 |
+
"logits/chosen": 0.12401044368743896,
|
| 1317 |
+
"logits/rejected": 0.11555816233158112,
|
| 1318 |
+
"logps/chosen": -461.777099609375,
|
| 1319 |
+
"logps/rejected": -446.89154052734375,
|
| 1320 |
+
"loss": 0.0958,
|
| 1321 |
+
"rewards/accuracies": 0.512499988079071,
|
| 1322 |
+
"rewards/chosen": -0.10517005622386932,
|
| 1323 |
+
"rewards/margins": 0.03241740167140961,
|
| 1324 |
+
"rewards/rejected": -0.13758745789527893,
|
| 1325 |
+
"step": 930
|
| 1326 |
+
},
|
| 1327 |
+
{
|
| 1328 |
+
"epoch": 0.59,
|
| 1329 |
+
"learning_rate": 2.1298565425660842e-06,
|
| 1330 |
+
"logits/chosen": 0.12125201523303986,
|
| 1331 |
+
"logits/rejected": 0.13117629289627075,
|
| 1332 |
+
"logps/chosen": -443.516845703125,
|
| 1333 |
+
"logps/rejected": -483.62030029296875,
|
| 1334 |
+
"loss": 0.086,
|
| 1335 |
+
"rewards/accuracies": 0.612500011920929,
|
| 1336 |
+
"rewards/chosen": -0.09896551072597504,
|
| 1337 |
+
"rewards/margins": 0.045016057789325714,
|
| 1338 |
+
"rewards/rejected": -0.14398157596588135,
|
| 1339 |
+
"step": 940
|
| 1340 |
+
},
|
| 1341 |
+
{
|
| 1342 |
+
"epoch": 0.6,
|
| 1343 |
+
"learning_rate": 2.0755188129062566e-06,
|
| 1344 |
+
"logits/chosen": 0.1026390790939331,
|
| 1345 |
+
"logits/rejected": 0.09158008545637131,
|
| 1346 |
+
"logps/chosen": -444.39862060546875,
|
| 1347 |
+
"logps/rejected": -479.50390625,
|
| 1348 |
+
"loss": 0.0983,
|
| 1349 |
+
"rewards/accuracies": 0.5562499761581421,
|
| 1350 |
+
"rewards/chosen": -0.10992302745580673,
|
| 1351 |
+
"rewards/margins": 0.03600381314754486,
|
| 1352 |
+
"rewards/rejected": -0.145926833152771,
|
| 1353 |
+
"step": 950
|
| 1354 |
+
},
|
| 1355 |
+
{
|
| 1356 |
+
"epoch": 0.61,
|
| 1357 |
+
"learning_rate": 2.0213868109153843e-06,
|
| 1358 |
+
"logits/chosen": 0.0887732207775116,
|
| 1359 |
+
"logits/rejected": 0.13315047323703766,
|
| 1360 |
+
"logps/chosen": -451.1080627441406,
|
| 1361 |
+
"logps/rejected": -477.95477294921875,
|
| 1362 |
+
"loss": 0.0883,
|
| 1363 |
+
"rewards/accuracies": 0.574999988079071,
|
| 1364 |
+
"rewards/chosen": -0.10277727991342545,
|
| 1365 |
+
"rewards/margins": 0.045780349522829056,
|
| 1366 |
+
"rewards/rejected": -0.1485576331615448,
|
| 1367 |
+
"step": 960
|
| 1368 |
+
},
|
| 1369 |
+
{
|
| 1370 |
+
"epoch": 0.61,
|
| 1371 |
+
"learning_rate": 1.967486772033164e-06,
|
| 1372 |
+
"logits/chosen": 0.1042795404791832,
|
| 1373 |
+
"logits/rejected": 0.1479608118534088,
|
| 1374 |
+
"logps/chosen": -433.328125,
|
| 1375 |
+
"logps/rejected": -472.20904541015625,
|
| 1376 |
+
"loss": 0.0802,
|
| 1377 |
+
"rewards/accuracies": 0.625,
|
| 1378 |
+
"rewards/chosen": -0.09643669426441193,
|
| 1379 |
+
"rewards/margins": 0.04762507602572441,
|
| 1380 |
+
"rewards/rejected": -0.14406177401542664,
|
| 1381 |
+
"step": 970
|
| 1382 |
+
},
|
| 1383 |
+
{
|
| 1384 |
+
"epoch": 0.62,
|
| 1385 |
+
"learning_rate": 1.9138448192768072e-06,
|
| 1386 |
+
"logits/chosen": 0.08118772506713867,
|
| 1387 |
+
"logits/rejected": 0.09627708047628403,
|
| 1388 |
+
"logps/chosen": -424.64898681640625,
|
| 1389 |
+
"logps/rejected": -493.44476318359375,
|
| 1390 |
+
"loss": 0.0919,
|
| 1391 |
+
"rewards/accuracies": 0.675000011920929,
|
| 1392 |
+
"rewards/chosen": -0.08744023740291595,
|
| 1393 |
+
"rewards/margins": 0.0591917559504509,
|
| 1394 |
+
"rewards/rejected": -0.14663198590278625,
|
| 1395 |
+
"step": 980
|
| 1396 |
+
},
|
| 1397 |
+
{
|
| 1398 |
+
"epoch": 0.62,
|
| 1399 |
+
"learning_rate": 1.8604869505803407e-06,
|
| 1400 |
+
"logits/chosen": 0.1322592794895172,
|
| 1401 |
+
"logits/rejected": 0.13863101601600647,
|
| 1402 |
+
"logps/chosen": -423.8280334472656,
|
| 1403 |
+
"logps/rejected": -469.0419921875,
|
| 1404 |
+
"loss": 0.0978,
|
| 1405 |
+
"rewards/accuracies": 0.643750011920929,
|
| 1406 |
+
"rewards/chosen": -0.09548251330852509,
|
| 1407 |
+
"rewards/margins": 0.05783646181225777,
|
| 1408 |
+
"rewards/rejected": -0.15331897139549255,
|
| 1409 |
+
"step": 990
|
| 1410 |
+
},
|
| 1411 |
+
{
|
| 1412 |
+
"epoch": 0.63,
|
| 1413 |
+
"learning_rate": 1.8074390261945418e-06,
|
| 1414 |
+
"logits/chosen": 0.10184427350759506,
|
| 1415 |
+
"logits/rejected": 0.12423571199178696,
|
| 1416 |
+
"logps/chosen": -377.3212585449219,
|
| 1417 |
+
"logps/rejected": -427.81317138671875,
|
| 1418 |
+
"loss": 0.0902,
|
| 1419 |
+
"rewards/accuracies": 0.53125,
|
| 1420 |
+
"rewards/chosen": -0.0946311205625534,
|
| 1421 |
+
"rewards/margins": 0.046779900789260864,
|
| 1422 |
+
"rewards/rejected": -0.14141102135181427,
|
| 1423 |
+
"step": 1000
|
| 1424 |
+
},
|
| 1425 |
+
{
|
| 1426 |
+
"epoch": 0.64,
|
| 1427 |
+
"learning_rate": 1.7547267561535836e-06,
|
| 1428 |
+
"logits/chosen": 0.06255786120891571,
|
| 1429 |
+
"logits/rejected": 0.12503619492053986,
|
| 1430 |
+
"logps/chosen": -418.79144287109375,
|
| 1431 |
+
"logps/rejected": -441.3435974121094,
|
| 1432 |
+
"loss": 0.0864,
|
| 1433 |
+
"rewards/accuracies": 0.612500011920929,
|
| 1434 |
+
"rewards/chosen": -0.10577164590358734,
|
| 1435 |
+
"rewards/margins": 0.05561785027384758,
|
| 1436 |
+
"rewards/rejected": -0.16138949990272522,
|
| 1437 |
+
"step": 1010
|
| 1438 |
+
},
|
| 1439 |
+
{
|
| 1440 |
+
"epoch": 0.64,
|
| 1441 |
+
"learning_rate": 1.702375687814501e-06,
|
| 1442 |
+
"logits/chosen": 0.10921992361545563,
|
| 1443 |
+
"logits/rejected": 0.0866740271449089,
|
| 1444 |
+
"logps/chosen": -404.1749267578125,
|
| 1445 |
+
"logps/rejected": -476.277099609375,
|
| 1446 |
+
"loss": 0.0791,
|
| 1447 |
+
"rewards/accuracies": 0.612500011920929,
|
| 1448 |
+
"rewards/chosen": -0.10695375502109528,
|
| 1449 |
+
"rewards/margins": 0.05870165675878525,
|
| 1450 |
+
"rewards/rejected": -0.16565541923046112,
|
| 1451 |
+
"step": 1020
|
| 1452 |
+
},
|
| 1453 |
+
{
|
| 1454 |
+
"epoch": 0.65,
|
| 1455 |
+
"learning_rate": 1.650411193475489e-06,
|
| 1456 |
+
"logits/chosen": 0.074369415640831,
|
| 1457 |
+
"logits/rejected": 0.12412852048873901,
|
| 1458 |
+
"logps/chosen": -420.69793701171875,
|
| 1459 |
+
"logps/rejected": -484.34161376953125,
|
| 1460 |
+
"loss": 0.0868,
|
| 1461 |
+
"rewards/accuracies": 0.643750011920929,
|
| 1462 |
+
"rewards/chosen": -0.11432509124279022,
|
| 1463 |
+
"rewards/margins": 0.051989488303661346,
|
| 1464 |
+
"rewards/rejected": -0.16631458699703217,
|
| 1465 |
+
"step": 1030
|
| 1466 |
+
},
|
| 1467 |
+
{
|
| 1468 |
+
"epoch": 0.66,
|
| 1469 |
+
"learning_rate": 1.5988584580790452e-06,
|
| 1470 |
+
"logits/chosen": 0.09442147612571716,
|
| 1471 |
+
"logits/rejected": 0.10659205913543701,
|
| 1472 |
+
"logps/chosen": -458.6270446777344,
|
| 1473 |
+
"logps/rejected": -523.6939697265625,
|
| 1474 |
+
"loss": 0.0705,
|
| 1475 |
+
"rewards/accuracies": 0.6000000238418579,
|
| 1476 |
+
"rewards/chosen": -0.10721229016780853,
|
| 1477 |
+
"rewards/margins": 0.05447040870785713,
|
| 1478 |
+
"rewards/rejected": -0.16168268024921417,
|
| 1479 |
+
"step": 1040
|
| 1480 |
+
},
|
| 1481 |
+
{
|
| 1482 |
+
"epoch": 0.66,
|
| 1483 |
+
"learning_rate": 1.5477424670059227e-06,
|
| 1484 |
+
"logits/chosen": 0.08504728972911835,
|
| 1485 |
+
"logits/rejected": 0.10464467108249664,
|
| 1486 |
+
"logps/chosen": -423.983154296875,
|
| 1487 |
+
"logps/rejected": -449.3802185058594,
|
| 1488 |
+
"loss": 0.0851,
|
| 1489 |
+
"rewards/accuracies": 0.581250011920929,
|
| 1490 |
+
"rewards/chosen": -0.11461503803730011,
|
| 1491 |
+
"rewards/margins": 0.03833803907036781,
|
| 1492 |
+
"rewards/rejected": -0.15295308828353882,
|
| 1493 |
+
"step": 1050
|
| 1494 |
+
},
|
| 1495 |
+
{
|
| 1496 |
+
"epoch": 0.67,
|
| 1497 |
+
"learning_rate": 1.4970879939657929e-06,
|
| 1498 |
+
"logits/chosen": 0.08304135501384735,
|
| 1499 |
+
"logits/rejected": 0.11797134578227997,
|
| 1500 |
+
"logps/chosen": -407.0821838378906,
|
| 1501 |
+
"logps/rejected": -456.46533203125,
|
| 1502 |
+
"loss": 0.0796,
|
| 1503 |
+
"rewards/accuracies": 0.643750011920929,
|
| 1504 |
+
"rewards/chosen": -0.08808410167694092,
|
| 1505 |
+
"rewards/margins": 0.05510552600026131,
|
| 1506 |
+
"rewards/rejected": -0.14318962395191193,
|
| 1507 |
+
"step": 1060
|
| 1508 |
+
},
|
| 1509 |
+
{
|
| 1510 |
+
"epoch": 0.67,
|
| 1511 |
+
"learning_rate": 1.4469195889904983e-06,
|
| 1512 |
+
"logits/chosen": 0.044432397931814194,
|
| 1513 |
+
"logits/rejected": 0.10810061544179916,
|
| 1514 |
+
"logps/chosen": -429.70794677734375,
|
| 1515 |
+
"logps/rejected": -452.116943359375,
|
| 1516 |
+
"loss": 0.0926,
|
| 1517 |
+
"rewards/accuracies": 0.6000000238418579,
|
| 1518 |
+
"rewards/chosen": -0.10013709962368011,
|
| 1519 |
+
"rewards/margins": 0.03565113991498947,
|
| 1520 |
+
"rewards/rejected": -0.13578826189041138,
|
| 1521 |
+
"step": 1070
|
| 1522 |
+
},
|
| 1523 |
+
{
|
| 1524 |
+
"epoch": 0.68,
|
| 1525 |
+
"learning_rate": 1.3972615665357228e-06,
|
| 1526 |
+
"logits/chosen": 0.08212409913539886,
|
| 1527 |
+
"logits/rejected": 0.08114130049943924,
|
| 1528 |
+
"logps/chosen": -425.5025939941406,
|
| 1529 |
+
"logps/rejected": -518.18115234375,
|
| 1530 |
+
"loss": 0.0884,
|
| 1531 |
+
"rewards/accuracies": 0.625,
|
| 1532 |
+
"rewards/chosen": -0.09838120639324188,
|
| 1533 |
+
"rewards/margins": 0.061208415776491165,
|
| 1534 |
+
"rewards/rejected": -0.15958961844444275,
|
| 1535 |
+
"step": 1080
|
| 1536 |
+
},
|
| 1537 |
+
{
|
| 1538 |
+
"epoch": 0.69,
|
| 1539 |
+
"learning_rate": 1.3481379936968154e-06,
|
| 1540 |
+
"logits/chosen": 0.06021580845117569,
|
| 1541 |
+
"logits/rejected": 0.1053982749581337,
|
| 1542 |
+
"logps/chosen": -471.2576599121094,
|
| 1543 |
+
"logps/rejected": -499.8750915527344,
|
| 1544 |
+
"loss": 0.1021,
|
| 1545 |
+
"rewards/accuracies": 0.5874999761581421,
|
| 1546 |
+
"rewards/chosen": -0.10752852261066437,
|
| 1547 |
+
"rewards/margins": 0.04095206782221794,
|
| 1548 |
+
"rewards/rejected": -0.1484805792570114,
|
| 1549 |
+
"step": 1090
|
| 1550 |
+
},
|
| 1551 |
+
{
|
| 1552 |
+
"epoch": 0.69,
|
| 1553 |
+
"learning_rate": 1.2995726785445187e-06,
|
| 1554 |
+
"logits/chosen": 0.08110491931438446,
|
| 1555 |
+
"logits/rejected": 0.13657256960868835,
|
| 1556 |
+
"logps/chosen": -457.90234375,
|
| 1557 |
+
"logps/rejected": -445.5087890625,
|
| 1558 |
+
"loss": 0.0918,
|
| 1559 |
+
"rewards/accuracies": 0.606249988079071,
|
| 1560 |
+
"rewards/chosen": -0.11181896924972534,
|
| 1561 |
+
"rewards/margins": 0.03636185824871063,
|
| 1562 |
+
"rewards/rejected": -0.14818084239959717,
|
| 1563 |
+
"step": 1100
|
| 1564 |
+
},
|
| 1565 |
+
{
|
| 1566 |
+
"epoch": 0.7,
|
| 1567 |
+
"learning_rate": 1.2515891585862248e-06,
|
| 1568 |
+
"logits/chosen": 0.06536366045475006,
|
| 1569 |
+
"logits/rejected": 0.10223551094532013,
|
| 1570 |
+
"logps/chosen": -424.5037536621094,
|
| 1571 |
+
"logps/rejected": -465.29425048828125,
|
| 1572 |
+
"loss": 0.112,
|
| 1573 |
+
"rewards/accuracies": 0.6000000238418579,
|
| 1574 |
+
"rewards/chosen": -0.10022538900375366,
|
| 1575 |
+
"rewards/margins": 0.03513466566801071,
|
| 1576 |
+
"rewards/rejected": -0.13536006212234497,
|
| 1577 |
+
"step": 1110
|
| 1578 |
+
},
|
| 1579 |
+
{
|
| 1580 |
+
"epoch": 0.71,
|
| 1581 |
+
"learning_rate": 1.204210689358359e-06,
|
| 1582 |
+
"logits/chosen": 0.15191073715686798,
|
| 1583 |
+
"logits/rejected": 0.12366287410259247,
|
| 1584 |
+
"logps/chosen": -436.87823486328125,
|
| 1585 |
+
"logps/rejected": -426.62762451171875,
|
| 1586 |
+
"loss": 0.0811,
|
| 1587 |
+
"rewards/accuracies": 0.5062500238418579,
|
| 1588 |
+
"rewards/chosen": -0.09861338883638382,
|
| 1589 |
+
"rewards/margins": 0.033628519624471664,
|
| 1590 |
+
"rewards/rejected": -0.13224190473556519,
|
| 1591 |
+
"step": 1120
|
| 1592 |
+
},
|
| 1593 |
+
{
|
| 1594 |
+
"epoch": 0.71,
|
| 1595 |
+
"learning_rate": 1.1574602331554352e-06,
|
| 1596 |
+
"logits/chosen": 0.05633612722158432,
|
| 1597 |
+
"logits/rejected": 0.04986396059393883,
|
| 1598 |
+
"logps/chosen": -424.77984619140625,
|
| 1599 |
+
"logps/rejected": -440.7940979003906,
|
| 1600 |
+
"loss": 0.1042,
|
| 1601 |
+
"rewards/accuracies": 0.5874999761581421,
|
| 1602 |
+
"rewards/chosen": -0.1184263676404953,
|
| 1603 |
+
"rewards/margins": 0.029354307800531387,
|
| 1604 |
+
"rewards/rejected": -0.1477806717157364,
|
| 1605 |
+
"step": 1130
|
| 1606 |
+
},
|
| 1607 |
+
{
|
| 1608 |
+
"epoch": 0.72,
|
| 1609 |
+
"learning_rate": 1.1113604479012202e-06,
|
| 1610 |
+
"logits/chosen": 0.04834729805588722,
|
| 1611 |
+
"logits/rejected": 0.079189732670784,
|
| 1612 |
+
"logps/chosen": -402.3204040527344,
|
| 1613 |
+
"logps/rejected": -482.1253967285156,
|
| 1614 |
+
"loss": 0.1046,
|
| 1615 |
+
"rewards/accuracies": 0.581250011920929,
|
| 1616 |
+
"rewards/chosen": -0.10266365110874176,
|
| 1617 |
+
"rewards/margins": 0.04556725174188614,
|
| 1618 |
+
"rewards/rejected": -0.1482308954000473,
|
| 1619 |
+
"step": 1140
|
| 1620 |
+
},
|
| 1621 |
+
{
|
| 1622 |
+
"epoch": 0.72,
|
| 1623 |
+
"learning_rate": 1.0659336761674163e-06,
|
| 1624 |
+
"logits/chosen": 0.07221159338951111,
|
| 1625 |
+
"logits/rejected": 0.06988708674907684,
|
| 1626 |
+
"logps/chosen": -378.05255126953125,
|
| 1627 |
+
"logps/rejected": -401.8747863769531,
|
| 1628 |
+
"loss": 0.0872,
|
| 1629 |
+
"rewards/accuracies": 0.581250011920929,
|
| 1630 |
+
"rewards/chosen": -0.09310795366764069,
|
| 1631 |
+
"rewards/margins": 0.04454265162348747,
|
| 1632 |
+
"rewards/rejected": -0.13765059411525726,
|
| 1633 |
+
"step": 1150
|
| 1634 |
+
},
|
| 1635 |
+
{
|
| 1636 |
+
"epoch": 0.73,
|
| 1637 |
+
"learning_rate": 1.0212019343451889e-06,
|
| 1638 |
+
"logits/chosen": 0.051744282245635986,
|
| 1639 |
+
"logits/rejected": 0.05535260960459709,
|
| 1640 |
+
"logps/chosen": -367.1768798828125,
|
| 1641 |
+
"logps/rejected": -443.02960205078125,
|
| 1642 |
+
"loss": 0.0823,
|
| 1643 |
+
"rewards/accuracies": 0.6187499761581421,
|
| 1644 |
+
"rewards/chosen": -0.08505652099847794,
|
| 1645 |
+
"rewards/margins": 0.06588119268417358,
|
| 1646 |
+
"rewards/rejected": -0.15093770623207092,
|
| 1647 |
+
"step": 1160
|
| 1648 |
+
},
|
| 1649 |
+
{
|
| 1650 |
+
"epoch": 0.74,
|
| 1651 |
+
"learning_rate": 9.771869019747708e-07,
|
| 1652 |
+
"logits/chosen": 0.0807739868760109,
|
| 1653 |
+
"logits/rejected": 0.09035824239253998,
|
| 1654 |
+
"logps/chosen": -437.5768127441406,
|
| 1655 |
+
"logps/rejected": -521.4659423828125,
|
| 1656 |
+
"loss": 0.0918,
|
| 1657 |
+
"rewards/accuracies": 0.637499988079071,
|
| 1658 |
+
"rewards/chosen": -0.11788342148065567,
|
| 1659 |
+
"rewards/margins": 0.07230274379253387,
|
| 1660 |
+
"rewards/rejected": -0.19018617272377014,
|
| 1661 |
+
"step": 1170
|
| 1662 |
+
},
|
| 1663 |
+
{
|
| 1664 |
+
"epoch": 0.74,
|
| 1665 |
+
"learning_rate": 9.339099112383263e-07,
|
| 1666 |
+
"logits/chosen": 0.08397279679775238,
|
| 1667 |
+
"logits/rejected": 0.12051109224557877,
|
| 1668 |
+
"logps/chosen": -429.1914978027344,
|
| 1669 |
+
"logps/rejected": -477.98968505859375,
|
| 1670 |
+
"loss": 0.0898,
|
| 1671 |
+
"rewards/accuracies": 0.643750011920929,
|
| 1672 |
+
"rewards/chosen": -0.1165337786078453,
|
| 1673 |
+
"rewards/margins": 0.0657678097486496,
|
| 1674 |
+
"rewards/rejected": -0.1823015958070755,
|
| 1675 |
+
"step": 1180
|
| 1676 |
+
},
|
| 1677 |
+
{
|
| 1678 |
+
"epoch": 0.75,
|
| 1679 |
+
"learning_rate": 8.913919366211679e-07,
|
| 1680 |
+
"logits/chosen": 0.01225881278514862,
|
| 1681 |
+
"logits/rejected": 0.08070291578769684,
|
| 1682 |
+
"logps/chosen": -453.7749938964844,
|
| 1683 |
+
"logps/rejected": -491.718017578125,
|
| 1684 |
+
"loss": 0.0784,
|
| 1685 |
+
"rewards/accuracies": 0.5874999761581421,
|
| 1686 |
+
"rewards/chosen": -0.12794212996959686,
|
| 1687 |
+
"rewards/margins": 0.046219903975725174,
|
| 1688 |
+
"rewards/rejected": -0.17416204512119293,
|
| 1689 |
+
"step": 1190
|
| 1690 |
+
},
|
| 1691 |
+
{
|
| 1692 |
+
"epoch": 0.76,
|
| 1693 |
+
"learning_rate": 8.496535847463264e-07,
|
| 1694 |
+
"logits/chosen": 0.05466824769973755,
|
| 1695 |
+
"logits/rejected": 0.06951786577701569,
|
| 1696 |
+
"logps/chosen": -409.43731689453125,
|
| 1697 |
+
"logps/rejected": -478.67730712890625,
|
| 1698 |
+
"loss": 0.0878,
|
| 1699 |
+
"rewards/accuracies": 0.637499988079071,
|
| 1700 |
+
"rewards/chosen": -0.0958373174071312,
|
| 1701 |
+
"rewards/margins": 0.06815023720264435,
|
| 1702 |
+
"rewards/rejected": -0.16398754715919495,
|
| 1703 |
+
"step": 1200
|
| 1704 |
+
},
|
| 1705 |
+
{
|
| 1706 |
+
"epoch": 0.76,
|
| 1707 |
+
"learning_rate": 8.08715084387417e-07,
|
| 1708 |
+
"logits/chosen": 0.0821441262960434,
|
| 1709 |
+
"logits/rejected": 0.09297174215316772,
|
| 1710 |
+
"logps/chosen": -443.9449768066406,
|
| 1711 |
+
"logps/rejected": -495.43304443359375,
|
| 1712 |
+
"loss": 0.0746,
|
| 1713 |
+
"rewards/accuracies": 0.6187499761581421,
|
| 1714 |
+
"rewards/chosen": -0.09165932238101959,
|
| 1715 |
+
"rewards/margins": 0.06980036199092865,
|
| 1716 |
+
"rewards/rejected": -0.16145966947078705,
|
| 1717 |
+
"step": 1210
|
| 1718 |
+
},
|
| 1719 |
+
{
|
| 1720 |
+
"epoch": 0.77,
|
| 1721 |
+
"learning_rate": 7.685962766646251e-07,
|
| 1722 |
+
"logits/chosen": 0.0996304377913475,
|
| 1723 |
+
"logits/rejected": 0.1416376680135727,
|
| 1724 |
+
"logps/chosen": -421.4239807128906,
|
| 1725 |
+
"logps/rejected": -448.3440856933594,
|
| 1726 |
+
"loss": 0.0854,
|
| 1727 |
+
"rewards/accuracies": 0.5874999761581421,
|
| 1728 |
+
"rewards/chosen": -0.1046370267868042,
|
| 1729 |
+
"rewards/margins": 0.04357679933309555,
|
| 1730 |
+
"rewards/rejected": -0.14821383357048035,
|
| 1731 |
+
"step": 1220
|
| 1732 |
+
},
|
| 1733 |
+
{
|
| 1734 |
+
"epoch": 0.78,
|
| 1735 |
+
"learning_rate": 7.29316605428573e-07,
|
| 1736 |
+
"logits/chosen": 0.050212837755680084,
|
| 1737 |
+
"logits/rejected": 0.08997879177331924,
|
| 1738 |
+
"logps/chosen": -405.666748046875,
|
| 1739 |
+
"logps/rejected": -481.518798828125,
|
| 1740 |
+
"loss": 0.0874,
|
| 1741 |
+
"rewards/accuracies": 0.668749988079071,
|
| 1742 |
+
"rewards/chosen": -0.10379258543252945,
|
| 1743 |
+
"rewards/margins": 0.06835626810789108,
|
| 1744 |
+
"rewards/rejected": -0.17214885354042053,
|
| 1745 |
+
"step": 1230
|
| 1746 |
+
},
|
| 1747 |
+
{
|
| 1748 |
+
"epoch": 0.78,
|
| 1749 |
+
"learning_rate": 6.908951078367284e-07,
|
| 1750 |
+
"logits/chosen": 0.06391054391860962,
|
| 1751 |
+
"logits/rejected": 0.11078351736068726,
|
| 1752 |
+
"logps/chosen": -472.7290954589844,
|
| 1753 |
+
"logps/rejected": -525.4988403320312,
|
| 1754 |
+
"loss": 0.0804,
|
| 1755 |
+
"rewards/accuracies": 0.643750011920929,
|
| 1756 |
+
"rewards/chosen": -0.11707256734371185,
|
| 1757 |
+
"rewards/margins": 0.06153646111488342,
|
| 1758 |
+
"rewards/rejected": -0.17860902845859528,
|
| 1759 |
+
"step": 1240
|
| 1760 |
+
},
|
| 1761 |
+
{
|
| 1762 |
+
"epoch": 0.79,
|
| 1763 |
+
"learning_rate": 6.533504051269185e-07,
|
| 1764 |
+
"logits/chosen": 0.09103193879127502,
|
| 1765 |
+
"logits/rejected": 0.1026821881532669,
|
| 1766 |
+
"logps/chosen": -356.58807373046875,
|
| 1767 |
+
"logps/rejected": -444.40869140625,
|
| 1768 |
+
"loss": 0.0813,
|
| 1769 |
+
"rewards/accuracies": 0.637499988079071,
|
| 1770 |
+
"rewards/chosen": -0.08095113933086395,
|
| 1771 |
+
"rewards/margins": 0.059402190148830414,
|
| 1772 |
+
"rewards/rejected": -0.14035332202911377,
|
| 1773 |
+
"step": 1250
|
| 1774 |
+
},
|
| 1775 |
+
{
|
| 1776 |
+
"epoch": 0.79,
|
| 1777 |
+
"learning_rate": 6.167006935924164e-07,
|
| 1778 |
+
"logits/chosen": 0.07466550171375275,
|
| 1779 |
+
"logits/rejected": 0.08906082808971405,
|
| 1780 |
+
"logps/chosen": -419.82733154296875,
|
| 1781 |
+
"logps/rejected": -447.23419189453125,
|
| 1782 |
+
"loss": 0.0858,
|
| 1783 |
+
"rewards/accuracies": 0.643750011920929,
|
| 1784 |
+
"rewards/chosen": -0.09368356317281723,
|
| 1785 |
+
"rewards/margins": 0.049560584127902985,
|
| 1786 |
+
"rewards/rejected": -0.14324414730072021,
|
| 1787 |
+
"step": 1260
|
| 1788 |
+
},
|
| 1789 |
+
{
|
| 1790 |
+
"epoch": 0.8,
|
| 1791 |
+
"learning_rate": 5.809637357629916e-07,
|
| 1792 |
+
"logits/chosen": 0.09017828106880188,
|
| 1793 |
+
"logits/rejected": 0.07519755512475967,
|
| 1794 |
+
"logps/chosen": -386.6206970214844,
|
| 1795 |
+
"logps/rejected": -444.78900146484375,
|
| 1796 |
+
"loss": 0.0884,
|
| 1797 |
+
"rewards/accuracies": 0.5687500238418579,
|
| 1798 |
+
"rewards/chosen": -0.09151438623666763,
|
| 1799 |
+
"rewards/margins": 0.0505540557205677,
|
| 1800 |
+
"rewards/rejected": -0.14206844568252563,
|
| 1801 |
+
"step": 1270
|
| 1802 |
+
},
|
| 1803 |
+
{
|
| 1804 |
+
"epoch": 0.81,
|
| 1805 |
+
"learning_rate": 5.461568517961763e-07,
|
| 1806 |
+
"logits/chosen": 0.022731300443410873,
|
| 1807 |
+
"logits/rejected": 0.03403092175722122,
|
| 1808 |
+
"logps/chosen": -406.1392822265625,
|
| 1809 |
+
"logps/rejected": -505.94366455078125,
|
| 1810 |
+
"loss": 0.0879,
|
| 1811 |
+
"rewards/accuracies": 0.6499999761581421,
|
| 1812 |
+
"rewards/chosen": -0.11663617938756943,
|
| 1813 |
+
"rewards/margins": 0.07096937298774719,
|
| 1814 |
+
"rewards/rejected": -0.18760555982589722,
|
| 1815 |
+
"step": 1280
|
| 1816 |
+
},
|
| 1817 |
+
{
|
| 1818 |
+
"epoch": 0.81,
|
| 1819 |
+
"learning_rate": 5.122969110829346e-07,
|
| 1820 |
+
"logits/chosen": 0.03181394562125206,
|
| 1821 |
+
"logits/rejected": 0.045416783541440964,
|
| 1822 |
+
"logps/chosen": -403.97100830078125,
|
| 1823 |
+
"logps/rejected": -524.0074462890625,
|
| 1824 |
+
"loss": 0.0895,
|
| 1825 |
+
"rewards/accuracies": 0.6625000238418579,
|
| 1826 |
+
"rewards/chosen": -0.11092919111251831,
|
| 1827 |
+
"rewards/margins": 0.07563652843236923,
|
| 1828 |
+
"rewards/rejected": -0.18656572699546814,
|
| 1829 |
+
"step": 1290
|
| 1830 |
+
},
|
| 1831 |
+
{
|
| 1832 |
+
"epoch": 0.82,
|
| 1833 |
+
"learning_rate": 4.794003240718004e-07,
|
| 1834 |
+
"logits/chosen": 0.09289757162332535,
|
| 1835 |
+
"logits/rejected": 0.09983552247285843,
|
| 1836 |
+
"logps/chosen": -467.2046813964844,
|
| 1837 |
+
"logps/rejected": -487.23760986328125,
|
| 1838 |
+
"loss": 0.095,
|
| 1839 |
+
"rewards/accuracies": 0.699999988079071,
|
| 1840 |
+
"rewards/chosen": -0.10705260187387466,
|
| 1841 |
+
"rewards/margins": 0.050529349595308304,
|
| 1842 |
+
"rewards/rejected": -0.15758195519447327,
|
| 1843 |
+
"step": 1300
|
| 1844 |
+
},
|
| 1845 |
+
{
|
| 1846 |
+
"epoch": 0.83,
|
| 1847 |
+
"learning_rate": 4.4748303431543907e-07,
|
| 1848 |
+
"logits/chosen": 0.07837380468845367,
|
| 1849 |
+
"logits/rejected": 0.10405907779932022,
|
| 1850 |
+
"logps/chosen": -479.838134765625,
|
| 1851 |
+
"logps/rejected": -476.2897033691406,
|
| 1852 |
+
"loss": 0.0852,
|
| 1853 |
+
"rewards/accuracies": 0.606249988079071,
|
| 1854 |
+
"rewards/chosen": -0.10937539488077164,
|
| 1855 |
+
"rewards/margins": 0.04146861657500267,
|
| 1856 |
+
"rewards/rejected": -0.150844007730484,
|
| 1857 |
+
"step": 1310
|
| 1858 |
+
},
|
| 1859 |
+
{
|
| 1860 |
+
"epoch": 0.83,
|
| 1861 |
+
"learning_rate": 4.1656051074350113e-07,
|
| 1862 |
+
"logits/chosen": 0.075148805975914,
|
| 1863 |
+
"logits/rejected": 0.08904428780078888,
|
| 1864 |
+
"logps/chosen": -405.589599609375,
|
| 1865 |
+
"logps/rejected": -448.927734375,
|
| 1866 |
+
"loss": 0.0901,
|
| 1867 |
+
"rewards/accuracies": 0.550000011920929,
|
| 1868 |
+
"rewards/chosen": -0.09542171657085419,
|
| 1869 |
+
"rewards/margins": 0.03852013498544693,
|
| 1870 |
+
"rewards/rejected": -0.1339418590068817,
|
| 1871 |
+
"step": 1320
|
| 1872 |
+
},
|
| 1873 |
+
{
|
| 1874 |
+
"epoch": 0.84,
|
| 1875 |
+
"learning_rate": 3.866477401655e-07,
|
| 1876 |
+
"logits/chosen": 0.07279117405414581,
|
| 1877 |
+
"logits/rejected": 0.08455020189285278,
|
| 1878 |
+
"logps/chosen": -385.3067626953125,
|
| 1879 |
+
"logps/rejected": -440.7181701660156,
|
| 1880 |
+
"loss": 0.0822,
|
| 1881 |
+
"rewards/accuracies": 0.5562499761581421,
|
| 1882 |
+
"rewards/chosen": -0.09135736525058746,
|
| 1883 |
+
"rewards/margins": 0.04145896062254906,
|
| 1884 |
+
"rewards/rejected": -0.13281632959842682,
|
| 1885 |
+
"step": 1330
|
| 1886 |
+
},
|
| 1887 |
+
{
|
| 1888 |
+
"epoch": 0.84,
|
| 1889 |
+
"learning_rate": 3.577592200073518e-07,
|
| 1890 |
+
"logits/chosen": 0.05959012359380722,
|
| 1891 |
+
"logits/rejected": 0.04779018834233284,
|
| 1892 |
+
"logps/chosen": -411.390625,
|
| 1893 |
+
"logps/rejected": -462.2435607910156,
|
| 1894 |
+
"loss": 0.0844,
|
| 1895 |
+
"rewards/accuracies": 0.625,
|
| 1896 |
+
"rewards/chosen": -0.09003390371799469,
|
| 1897 |
+
"rewards/margins": 0.03744681924581528,
|
| 1898 |
+
"rewards/rejected": -0.12748071551322937,
|
| 1899 |
+
"step": 1340
|
| 1900 |
+
},
|
| 1901 |
+
{
|
| 1902 |
+
"epoch": 0.85,
|
| 1903 |
+
"learning_rate": 3.2990895128510364e-07,
|
| 1904 |
+
"logits/chosen": 0.05087254196405411,
|
| 1905 |
+
"logits/rejected": 0.07081006467342377,
|
| 1906 |
+
"logps/chosen": -413.0889587402344,
|
| 1907 |
+
"logps/rejected": -446.05242919921875,
|
| 1908 |
+
"loss": 0.0881,
|
| 1909 |
+
"rewards/accuracies": 0.6312500238418579,
|
| 1910 |
+
"rewards/chosen": -0.09529262781143188,
|
| 1911 |
+
"rewards/margins": 0.045550812035799026,
|
| 1912 |
+
"rewards/rejected": -0.1408434361219406,
|
| 1913 |
+
"step": 1350
|
| 1914 |
+
},
|
| 1915 |
+
{
|
| 1916 |
+
"epoch": 0.86,
|
| 1917 |
+
"learning_rate": 3.0311043181924237e-07,
|
| 1918 |
+
"logits/chosen": 0.028466597199440002,
|
| 1919 |
+
"logits/rejected": 0.06887885928153992,
|
| 1920 |
+
"logps/chosen": -412.6106872558594,
|
| 1921 |
+
"logps/rejected": -454.3301696777344,
|
| 1922 |
+
"loss": 0.0969,
|
| 1923 |
+
"rewards/accuracies": 0.59375,
|
| 1924 |
+
"rewards/chosen": -0.10608142614364624,
|
| 1925 |
+
"rewards/margins": 0.02706090547144413,
|
| 1926 |
+
"rewards/rejected": -0.13314232230186462,
|
| 1927 |
+
"step": 1360
|
| 1928 |
+
},
|
| 1929 |
+
{
|
| 1930 |
+
"epoch": 0.86,
|
| 1931 |
+
"learning_rate": 2.773766496928898e-07,
|
| 1932 |
+
"logits/chosen": 0.06292058527469635,
|
| 1933 |
+
"logits/rejected": 0.08972954005002975,
|
| 1934 |
+
"logps/chosen": -411.333251953125,
|
| 1935 |
+
"logps/rejected": -474.076416015625,
|
| 1936 |
+
"loss": 0.0967,
|
| 1937 |
+
"rewards/accuracies": 0.6187499761581421,
|
| 1938 |
+
"rewards/chosen": -0.0901075154542923,
|
| 1939 |
+
"rewards/margins": 0.06678055226802826,
|
| 1940 |
+
"rewards/rejected": -0.15688805282115936,
|
| 1941 |
+
"step": 1370
|
| 1942 |
+
},
|
| 1943 |
+
{
|
| 1944 |
+
"epoch": 0.87,
|
| 1945 |
+
"learning_rate": 2.527200769570376e-07,
|
| 1946 |
+
"logits/chosen": 0.05736211687326431,
|
| 1947 |
+
"logits/rejected": 0.047862760722637177,
|
| 1948 |
+
"logps/chosen": -428.587890625,
|
| 1949 |
+
"logps/rejected": -487.86029052734375,
|
| 1950 |
+
"loss": 0.091,
|
| 1951 |
+
"rewards/accuracies": 0.612500011920929,
|
| 1952 |
+
"rewards/chosen": -0.11271530389785767,
|
| 1953 |
+
"rewards/margins": 0.05474567413330078,
|
| 1954 |
+
"rewards/rejected": -0.16746096312999725,
|
| 1955 |
+
"step": 1380
|
| 1956 |
+
},
|
| 1957 |
+
{
|
| 1958 |
+
"epoch": 0.88,
|
| 1959 |
+
"learning_rate": 2.291526635858829e-07,
|
| 1960 |
+
"logits/chosen": 0.033168576657772064,
|
| 1961 |
+
"logits/rejected": 0.07321670651435852,
|
| 1962 |
+
"logps/chosen": -406.710205078125,
|
| 1963 |
+
"logps/rejected": -515.9356079101562,
|
| 1964 |
+
"loss": 0.0829,
|
| 1965 |
+
"rewards/accuracies": 0.65625,
|
| 1966 |
+
"rewards/chosen": -0.09568953514099121,
|
| 1967 |
+
"rewards/margins": 0.058294862508773804,
|
| 1968 |
+
"rewards/rejected": -0.1539844125509262,
|
| 1969 |
+
"step": 1390
|
| 1970 |
+
},
|
| 1971 |
+
{
|
| 1972 |
+
"epoch": 0.88,
|
| 1973 |
+
"learning_rate": 2.0668583168519608e-07,
|
| 1974 |
+
"logits/chosen": 0.047509074211120605,
|
| 1975 |
+
"logits/rejected": 0.06223832815885544,
|
| 1976 |
+
"logps/chosen": -388.98193359375,
|
| 1977 |
+
"logps/rejected": -471.21136474609375,
|
| 1978 |
+
"loss": 0.0839,
|
| 1979 |
+
"rewards/accuracies": 0.59375,
|
| 1980 |
+
"rewards/chosen": -0.09332209825515747,
|
| 1981 |
+
"rewards/margins": 0.04638640210032463,
|
| 1982 |
+
"rewards/rejected": -0.1397085040807724,
|
| 1983 |
+
"step": 1400
|
| 1984 |
+
},
|
| 1985 |
+
{
|
| 1986 |
+
"epoch": 0.89,
|
| 1987 |
+
"learning_rate": 1.8533046995651755e-07,
|
| 1988 |
+
"logits/chosen": 0.06117665767669678,
|
| 1989 |
+
"logits/rejected": 0.05806068703532219,
|
| 1990 |
+
"logps/chosen": -398.0728454589844,
|
| 1991 |
+
"logps/rejected": -476.2325744628906,
|
| 1992 |
+
"loss": 0.0928,
|
| 1993 |
+
"rewards/accuracies": 0.6312500238418579,
|
| 1994 |
+
"rewards/chosen": -0.10759732872247696,
|
| 1995 |
+
"rewards/margins": 0.06671231985092163,
|
| 1996 |
+
"rewards/rejected": -0.1743096262216568,
|
| 1997 |
+
"step": 1410
|
| 1998 |
+
},
|
| 1999 |
+
{
|
| 2000 |
+
"epoch": 0.9,
|
| 2001 |
+
"learning_rate": 1.6509692841987557e-07,
|
| 2002 |
+
"logits/chosen": 0.09554895013570786,
|
| 2003 |
+
"logits/rejected": 0.11183027923107147,
|
| 2004 |
+
"logps/chosen": -373.590576171875,
|
| 2005 |
+
"logps/rejected": -459.16046142578125,
|
| 2006 |
+
"loss": 0.0958,
|
| 2007 |
+
"rewards/accuracies": 0.65625,
|
| 2008 |
+
"rewards/chosen": -0.09168444573879242,
|
| 2009 |
+
"rewards/margins": 0.05873497575521469,
|
| 2010 |
+
"rewards/rejected": -0.1504194289445877,
|
| 2011 |
+
"step": 1420
|
| 2012 |
+
},
|
| 2013 |
+
{
|
| 2014 |
+
"epoch": 0.9,
|
| 2015 |
+
"learning_rate": 1.4599501339758043e-07,
|
| 2016 |
+
"logits/chosen": 0.14753518998622894,
|
| 2017 |
+
"logits/rejected": 0.15233811736106873,
|
| 2018 |
+
"logps/chosen": -391.5714416503906,
|
| 2019 |
+
"logps/rejected": -454.29901123046875,
|
| 2020 |
+
"loss": 0.0838,
|
| 2021 |
+
"rewards/accuracies": 0.65625,
|
| 2022 |
+
"rewards/chosen": -0.07346367090940475,
|
| 2023 |
+
"rewards/margins": 0.0692339688539505,
|
| 2024 |
+
"rewards/rejected": -0.14269763231277466,
|
| 2025 |
+
"step": 1430
|
| 2026 |
+
},
|
| 2027 |
+
{
|
| 2028 |
+
"epoch": 0.91,
|
| 2029 |
+
"learning_rate": 1.2803398276152236e-07,
|
| 2030 |
+
"logits/chosen": 0.0984094962477684,
|
| 2031 |
+
"logits/rejected": 0.158823162317276,
|
| 2032 |
+
"logps/chosen": -421.304931640625,
|
| 2033 |
+
"logps/rejected": -486.080810546875,
|
| 2034 |
+
"loss": 0.0911,
|
| 2035 |
+
"rewards/accuracies": 0.637499988079071,
|
| 2036 |
+
"rewards/chosen": -0.10145977884531021,
|
| 2037 |
+
"rewards/margins": 0.05414135009050369,
|
| 2038 |
+
"rewards/rejected": -0.1556011140346527,
|
| 2039 |
+
"step": 1440
|
| 2040 |
+
},
|
| 2041 |
+
{
|
| 2042 |
+
"epoch": 0.91,
|
| 2043 |
+
"learning_rate": 1.1122254144628308e-07,
|
| 2044 |
+
"logits/chosen": 0.03182172030210495,
|
| 2045 |
+
"logits/rejected": 0.06180509179830551,
|
| 2046 |
+
"logps/chosen": -370.4424743652344,
|
| 2047 |
+
"logps/rejected": -450.4090881347656,
|
| 2048 |
+
"loss": 0.0875,
|
| 2049 |
+
"rewards/accuracies": 0.5874999761581421,
|
| 2050 |
+
"rewards/chosen": -0.09365768730640411,
|
| 2051 |
+
"rewards/margins": 0.041229672729969025,
|
| 2052 |
+
"rewards/rejected": -0.13488735258579254,
|
| 2053 |
+
"step": 1450
|
| 2054 |
+
},
|
| 2055 |
+
{
|
| 2056 |
+
"epoch": 0.92,
|
| 2057 |
+
"learning_rate": 9.556883723023047e-08,
|
| 2058 |
+
"logits/chosen": 0.03198453038930893,
|
| 2059 |
+
"logits/rejected": 0.09605802595615387,
|
| 2060 |
+
"logps/chosen": -458.57366943359375,
|
| 2061 |
+
"logps/rejected": -517.7696533203125,
|
| 2062 |
+
"loss": 0.0928,
|
| 2063 |
+
"rewards/accuracies": 0.6625000238418579,
|
| 2064 |
+
"rewards/chosen": -0.10384533554315567,
|
| 2065 |
+
"rewards/margins": 0.06892184168100357,
|
| 2066 |
+
"rewards/rejected": -0.17276716232299805,
|
| 2067 |
+
"step": 1460
|
| 2068 |
+
},
|
| 2069 |
+
{
|
| 2070 |
+
"epoch": 0.93,
|
| 2071 |
+
"learning_rate": 8.108045678664133e-08,
|
| 2072 |
+
"logits/chosen": 0.046630989760160446,
|
| 2073 |
+
"logits/rejected": 0.11219167709350586,
|
| 2074 |
+
"logps/chosen": -429.394287109375,
|
| 2075 |
+
"logps/rejected": -471.2891540527344,
|
| 2076 |
+
"loss": 0.0851,
|
| 2077 |
+
"rewards/accuracies": 0.5874999761581421,
|
| 2078 |
+
"rewards/chosen": -0.10102847963571548,
|
| 2079 |
+
"rewards/margins": 0.04838350787758827,
|
| 2080 |
+
"rewards/rejected": -0.14941199123859406,
|
| 2081 |
+
"step": 1470
|
| 2082 |
+
},
|
| 2083 |
+
{
|
| 2084 |
+
"epoch": 0.93,
|
| 2085 |
+
"learning_rate": 6.776442200677136e-08,
|
| 2086 |
+
"logits/chosen": 0.07663191854953766,
|
| 2087 |
+
"logits/rejected": 0.08462060987949371,
|
| 2088 |
+
"logps/chosen": -418.231201171875,
|
| 2089 |
+
"logps/rejected": -460.2142639160156,
|
| 2090 |
+
"loss": 0.0978,
|
| 2091 |
+
"rewards/accuracies": 0.6000000238418579,
|
| 2092 |
+
"rewards/chosen": -0.08978335559368134,
|
| 2093 |
+
"rewards/margins": 0.05147411674261093,
|
| 2094 |
+
"rewards/rejected": -0.14125746488571167,
|
| 2095 |
+
"step": 1480
|
| 2096 |
+
},
|
| 2097 |
+
{
|
| 2098 |
+
"epoch": 0.94,
|
| 2099 |
+
"learning_rate": 5.562718659664651e-08,
|
| 2100 |
+
"logits/chosen": 0.05606203153729439,
|
| 2101 |
+
"logits/rejected": 0.03971995785832405,
|
| 2102 |
+
"logps/chosen": -446.9505920410156,
|
| 2103 |
+
"logps/rejected": -492.7408752441406,
|
| 2104 |
+
"loss": 0.1003,
|
| 2105 |
+
"rewards/accuracies": 0.643750011920929,
|
| 2106 |
+
"rewards/chosen": -0.11127449572086334,
|
| 2107 |
+
"rewards/margins": 0.049640461802482605,
|
| 2108 |
+
"rewards/rejected": -0.16091494262218475,
|
| 2109 |
+
"step": 1490
|
| 2110 |
+
},
|
| 2111 |
+
{
|
| 2112 |
+
"epoch": 0.95,
|
| 2113 |
+
"learning_rate": 4.467463294923091e-08,
|
| 2114 |
+
"logits/chosen": 0.07213045656681061,
|
| 2115 |
+
"logits/rejected": 0.11403784900903702,
|
| 2116 |
+
"logps/chosen": -408.98272705078125,
|
| 2117 |
+
"logps/rejected": -505.5140686035156,
|
| 2118 |
+
"loss": 0.0818,
|
| 2119 |
+
"rewards/accuracies": 0.643750011920929,
|
| 2120 |
+
"rewards/chosen": -0.10484303534030914,
|
| 2121 |
+
"rewards/margins": 0.07243579626083374,
|
| 2122 |
+
"rewards/rejected": -0.17727883160114288,
|
| 2123 |
+
"step": 1500
|
| 2124 |
+
},
|
| 2125 |
+
{
|
| 2126 |
+
"epoch": 0.95,
|
| 2127 |
+
"learning_rate": 3.491206929348595e-08,
|
| 2128 |
+
"logits/chosen": 0.09976326674222946,
|
| 2129 |
+
"logits/rejected": 0.11376956850290298,
|
| 2130 |
+
"logps/chosen": -462.38946533203125,
|
| 2131 |
+
"logps/rejected": -540.771484375,
|
| 2132 |
+
"loss": 0.083,
|
| 2133 |
+
"rewards/accuracies": 0.6625000238418579,
|
| 2134 |
+
"rewards/chosen": -0.11446479707956314,
|
| 2135 |
+
"rewards/margins": 0.06572078168392181,
|
| 2136 |
+
"rewards/rejected": -0.18018554151058197,
|
| 2137 |
+
"step": 1510
|
| 2138 |
+
},
|
| 2139 |
+
{
|
| 2140 |
+
"epoch": 0.96,
|
| 2141 |
+
"learning_rate": 2.634422712170209e-08,
|
| 2142 |
+
"logits/chosen": 0.06767505407333374,
|
| 2143 |
+
"logits/rejected": 0.08407372236251831,
|
| 2144 |
+
"logps/chosen": -445.62646484375,
|
| 2145 |
+
"logps/rejected": -456.46685791015625,
|
| 2146 |
+
"loss": 0.0938,
|
| 2147 |
+
"rewards/accuracies": 0.5562499761581421,
|
| 2148 |
+
"rewards/chosen": -0.11092151701450348,
|
| 2149 |
+
"rewards/margins": 0.028619635850191116,
|
| 2150 |
+
"rewards/rejected": -0.1395411491394043,
|
| 2151 |
+
"step": 1520
|
| 2152 |
+
},
|
| 2153 |
+
{
|
| 2154 |
+
"epoch": 0.96,
|
| 2155 |
+
"learning_rate": 1.8975258896347383e-08,
|
| 2156 |
+
"logits/chosen": 0.03723181039094925,
|
| 2157 |
+
"logits/rejected": 0.07889989018440247,
|
| 2158 |
+
"logps/chosen": -400.19281005859375,
|
| 2159 |
+
"logps/rejected": -445.9078063964844,
|
| 2160 |
+
"loss": 0.1039,
|
| 2161 |
+
"rewards/accuracies": 0.581250011920929,
|
| 2162 |
+
"rewards/chosen": -0.09558527171611786,
|
| 2163 |
+
"rewards/margins": 0.049675554037094116,
|
| 2164 |
+
"rewards/rejected": -0.14526084065437317,
|
| 2165 |
+
"step": 1530
|
| 2166 |
+
},
|
| 2167 |
+
{
|
| 2168 |
+
"epoch": 0.97,
|
| 2169 |
+
"learning_rate": 1.2808736037550618e-08,
|
| 2170 |
+
"logits/chosen": 0.022983688861131668,
|
| 2171 |
+
"logits/rejected": 0.07783511281013489,
|
| 2172 |
+
"logps/chosen": -435.6316833496094,
|
| 2173 |
+
"logps/rejected": -445.15118408203125,
|
| 2174 |
+
"loss": 0.0877,
|
| 2175 |
+
"rewards/accuracies": 0.5874999761581421,
|
| 2176 |
+
"rewards/chosen": -0.104395292699337,
|
| 2177 |
+
"rewards/margins": 0.03368845209479332,
|
| 2178 |
+
"rewards/rejected": -0.13808374106884003,
|
| 2179 |
+
"step": 1540
|
| 2180 |
+
},
|
| 2181 |
+
{
|
| 2182 |
+
"epoch": 0.98,
|
| 2183 |
+
"learning_rate": 7.847647192186226e-09,
|
| 2184 |
+
"logits/chosen": 0.08826269954442978,
|
| 2185 |
+
"logits/rejected": 0.08017078787088394,
|
| 2186 |
+
"logps/chosen": -440.181884765625,
|
| 2187 |
+
"logps/rejected": -492.92510986328125,
|
| 2188 |
+
"loss": 0.0749,
|
| 2189 |
+
"rewards/accuracies": 0.65625,
|
| 2190 |
+
"rewards/chosen": -0.09225556254386902,
|
| 2191 |
+
"rewards/margins": 0.06564488261938095,
|
| 2192 |
+
"rewards/rejected": -0.15790043771266937,
|
| 2193 |
+
"step": 1550
|
| 2194 |
+
},
|
| 2195 |
+
{
|
| 2196 |
+
"epoch": 0.98,
|
| 2197 |
+
"learning_rate": 4.094396785408461e-09,
|
| 2198 |
+
"logits/chosen": 0.050535012036561966,
|
| 2199 |
+
"logits/rejected": 0.09702993929386139,
|
| 2200 |
+
"logps/chosen": -396.10906982421875,
|
| 2201 |
+
"logps/rejected": -434.914794921875,
|
| 2202 |
+
"loss": 0.0964,
|
| 2203 |
+
"rewards/accuracies": 0.5874999761581421,
|
| 2204 |
+
"rewards/chosen": -0.09921000152826309,
|
| 2205 |
+
"rewards/margins": 0.04018466919660568,
|
| 2206 |
+
"rewards/rejected": -0.13939465582370758,
|
| 2207 |
+
"step": 1560
|
| 2208 |
+
},
|
| 2209 |
+
{
|
| 2210 |
+
"epoch": 0.99,
|
| 2211 |
+
"learning_rate": 1.5508038553291304e-09,
|
| 2212 |
+
"logits/chosen": 0.06177855283021927,
|
| 2213 |
+
"logits/rejected": 0.08348623663187027,
|
| 2214 |
+
"logps/chosen": -436.14166259765625,
|
| 2215 |
+
"logps/rejected": -509.82373046875,
|
| 2216 |
+
"loss": 0.0862,
|
| 2217 |
+
"rewards/accuracies": 0.6187499761581421,
|
| 2218 |
+
"rewards/chosen": -0.09413688629865646,
|
| 2219 |
+
"rewards/margins": 0.05878259614109993,
|
| 2220 |
+
"rewards/rejected": -0.15291950106620789,
|
| 2221 |
+
"step": 1570
|
| 2222 |
+
},
|
| 2223 |
+
{
|
| 2224 |
+
"epoch": 1.0,
|
| 2225 |
+
"learning_rate": 2.1810117140974584e-10,
|
| 2226 |
+
"logits/chosen": 0.024281805381178856,
|
| 2227 |
+
"logits/rejected": 0.09278794378042221,
|
| 2228 |
+
"logps/chosen": -491.00299072265625,
|
| 2229 |
+
"logps/rejected": -472.14349365234375,
|
| 2230 |
+
"loss": 0.1018,
|
| 2231 |
+
"rewards/accuracies": 0.6187499761581421,
|
| 2232 |
+
"rewards/chosen": -0.11561523377895355,
|
| 2233 |
+
"rewards/margins": 0.04595177248120308,
|
| 2234 |
+
"rewards/rejected": -0.16156700253486633,
|
| 2235 |
+
"step": 1580
|
| 2236 |
+
},
|
| 2237 |
+
{
|
| 2238 |
+
"epoch": 1.0,
|
| 2239 |
+
"step": 1586,
|
| 2240 |
+
"total_flos": 0.0,
|
| 2241 |
+
"train_loss": 0.09621614837856943,
|
| 2242 |
+
"train_runtime": 21999.9134,
|
| 2243 |
+
"train_samples_per_second": 3.461,
|
| 2244 |
+
"train_steps_per_second": 0.072
|
| 2245 |
+
}
|
| 2246 |
+
],
|
| 2247 |
+
"logging_steps": 10,
|
| 2248 |
+
"max_steps": 1586,
|
| 2249 |
+
"num_input_tokens_seen": 0,
|
| 2250 |
+
"num_train_epochs": 1,
|
| 2251 |
+
"save_steps": 100,
|
| 2252 |
+
"total_flos": 0.0,
|
| 2253 |
+
"train_batch_size": 4,
|
| 2254 |
+
"trial_name": null,
|
| 2255 |
+
"trial_params": null
|
| 2256 |
+
}
|